Pipeline Safety Trust | website for The Pipeline Safety Trustpstrust.org/docs/csfm_doc1.pdf · We...

NIPERIBDM-0357 HilifiOKlAHOMA A BOM Federal Company National Institute for Petroleum and Energy ResearchPost Office Box 2565Bartlesville, OK 74005 CALIFORNIA STATE FIRE MARSHAL RISK ASSESSMENT OF CALIFORNIA LOW PRESSURE CRUDE OIL AND CRUDE OIL GATHERING LINES for Management and Operating Contract for the Department of Energy's National Oil and Related Programs by EDM Services, Inc. Under Contract to BDM-Oklahoma, Inc. J. Stephen Jones, Program Manager July 1996 Work Performed Under Contract No. DE-AC22-94PC91008 Prepared for US. Depar~ent of Energy National Petroleum Technology Office

National Institute for Dill Petroleum and _______________ Energy Research oaAIIo~vfA. ost ce X A BOM rederal Company Bardesville, OK 74005 Phone 918 336-2400 Fax 918 337-4365 Direct Dial: 918 337-4458 E-mail: [email protected] Per Your Request: BDM-Oklahoma, Inc., an agent for the U. S. Department of Energy, National Petroleum Technology Office formerly the Bartlesville Project Office, is pleased to send you the enclosed information. Thank you for your interest in our research programs. I would like to invite you to our Oil Program Home Page address at: http://www.npto.doe.gov The World Wide Web Site contains information on the latest publications, events, seminars, workshops, and symposiums in the industry along with updates on current research programs and demonstration projects. We created this Web Site to provide easily accessible, accurate and timely information needed for your business. We look forward to hearing from you! 8DM-Oklahoma is the Management and Operating Contractor for the U. S. Department of Energy National Oil Program and the National Institute for Petroleum and Energy Research NIPER facility in Bartlesville, Oklahoma. If you have questions, please contact me at 918 337-4458, or at my E-mail address listed above. Sincerely, BDM-OKLAHOMA, INC. ~ Dorothyann Olson Computer Specialist! Technology Transfer Specialist

NIPERJBDM-0357 July 1996 CALIFORNIA STATE FIRE MARSHALRISK ASSESSMENT OF CALIFORNIA LOW PRESSURE CRUDE OIL AND CRUDE OIL GATHERING LINES for Management and Operating Contract for the Department of Energy's National Oil and Related Programs Work Performed Under Contract No. DE-AC22-94PC91008 Prepared for US. Department of Energy National Petroleum Technology Office DISCLAIMER This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. BDM-Oklahoma, [nc. P.O. Box 2565 Bartlesville, Oklahoma 74005

California State Fire Marshal Pip~line Assessment Steering CommitteeNancy Wolfe Jerry SimmonsDivision Chief, Pipeline Safety and Enforcement Program ManagerCalifornia State Fire Marshal BDM OklahomaTom Berg Les ClarkDirector, Resource Management Agency Vice PresidentVentura County Independent Oil Producers AgencyJohn Donovan John EuphratDirector, Environmental and Regulatory Affairs Principal PlannerCalifornia Independent Petroleum ~A~ssociation San Luis Obispo CountyBill Guerard Frank HolmesState Oil and Gas Supervisor Coastal CoordinatorCalifornia Division of Oil, Gas and Geothermal Resources Western States Petroleum AssociationCraig Jackson Nathan ManskeEnvironmental Regulatory Compliance Coordinator Lobbyist for Advocation and ResearchTexaco, U.S.A. KAHL AssociatesBarry McMahan Dan MilhalikAssistant Vice President Operations CoordinatorSeneca Resources Corporation Texaco Trading and Transportation, Inc.Mike Niblett Jim NorrisPetroleum Specialist Petroleum CoordinatorPetroleum Department, Santa Barbara County Building and Safety, Santa Barbara CountyCathy Reheis Ralph WarnngtonManaging Coordinator Senior Staff EngineerWestern States Petroleum Associabon Cal Resources, LLC NoticeThis document was prepared by EDM Services, Inc., under contract to BDM Oklahoma. All data wasfurnished by the pipeline operators~. EDM Services, Inc., BDM Oklahoma, United States Department ofEnergy, and/or the California State Fire Marshal, and their staffs do not: warrant the accuracy or completeness of the data collected, nor assume any liability resulting from the use of, or damage resulting from any information present~d herein.EDM Services, Inc., BDM Oklahoma, United States Department of Energy, and/or the California State FireMarshal do not endorse products o~ manufacturers. Trade or manufacturers' names appear herein solelybecause they are considered essen~ial to the objectives of this study.

California State Fire Marshal ~ July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STATBnR~~~AL California State Fire Marshal California Low Pressure Crude Oil and Crude Oil Gathering Lines ContentsExecutive Summary....................................................................................81.0 Introduction...................................................................................13 1.1 Regulatory Authority.....................................................................14 1.2 Relative Safety Perspective..............................................................15 1.3 Acknowledgments..........................................................................192.0 Methodology...................................................................................20 2.1 Funding and Contracting..................................................................20 2.2 Steering Committee.......................................................................20 2.3 Identify Study Participants and Pipelines................................................21 2.4 Data Gathering...........................................................................22 2.5 Database Development.....................................................................23 2.6 Field Audits.............................................................................23 2.7 Barriers and Incentive Options...........................................................24 2.8 Potential Data Inconsistencies...........................................................253.0 Background Pipeline Risk Data.................................................................26 3.1 CONCAWE- 1981 Through 1994...............................................................26 3.2 U.S. Natural Gas Transmission and Gathering Lines 1970 Through June 1984................................................28 3.3 U.S. Natural Gas Transmission and Gathering Lines June 1984 through 1992................................................30 3.4 U.S. Hazardous Liquid Pipeline Accidents 1986 through 1992.....................................................32 3.5 Data Summary of Regulated California Hazardous Liquid Pipelines 1981 through 1990.....................................................34 3.6 Data Summary of California Crude Oil Pipelines Under Study 1993 through 1995.....................................................34 3.7 Comparison of Various Incident Data Sources..............................................38 3.8 Uncorrected Pipeline Risks...............................................................414.0 General Risk Levels...........................................................................42 4.1 Overall Incident Causes..................................................................45 4.2 Incident Rates By Study Year.............................................................48 3

California State Fire Marshal July 1996STAflFIREMAi~SHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gatherin2 Lines 4.3 Decade of Con~truction Effects...............................52 4.4 Operating Temperature Effects................................56 4.5 Pipe Diameter Effects........................................59 4.6 Leak Detection~Systems.......................................64 4.7 Cathodic Protection System...................................68 4.8 Pipe Specification Effects...................................72 4.9 Pipe Type Effe~ts............................................78 4.10 Operating Pressure Effects...................................83 4.11 External Pipe Coatings.......................................87 4.12 Internal Inspect~ons.........................................92 4.13 Seasonal Effects.............................................94 4.14 Pipeline Components..........................................97 4.15 Hydrostatic Testing Interval.................................97 4.16 Spill Size Distribution.....................................105 4.17 Damage Distribution.........................................112 4.18 Incident Rates b' Internal Coating or Lining...............117 4.19 Incident Rates by Above versus Below Grade Pipe.............117 4.20 Recovery of Spi~led Volumes.................................117 4.21 Injuries and Fatalities.....................................1195.0 Barriers and Incentive Options..........................................121 5.1 Summary of Questionnaire Results: Incentives................122 5.2 Incentive Implementation....................................124 5.3 Summary of QuÁstionnaire Results: Barriers..................124 5.4 Actual and Poter~tial Consequences of Barriers..............126 5.5 Removing Barriers...........................................126 5.6 Case Studies................................................1286.0 Conclusions.............................................................132 6.1 Database and Sti.~dy Findings...............................132 6.2 Incentive Option~ Investigation Findings....................1377.0 Recommendations.........................................................139 7.1 Database and StUdy..........................................139 7.2 Barriers and Incentive Options..............................1408.0 Bibliography............................................................142 4

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~!R~MM~SHAL List of TablesTable 1-1 Fatalities by Mode of Transportation.................................................17Table 1-2 Estimated Fatalities Associated with Revenue Freight by Mode of Transportation............................................18Table 1-3 Estimated 1988 Fatalities per Billion Ton-Miles Transported by Mode of Transportation................................................18Table 3-1 European Hazardous Liquid Pipeline Incidents.........................................27Table 3-2 U. S. Natural Gas Transmission and Gathering Lines 1970 through June 1984...............................................................29Table 3-3 Onshore U. S. Natural Gas Transmission and Gathering Lines June 1984 through 1992...............................................................31Table 3-4 U. S. Hazardous Liquid Pipeline Accidents 1986 through 1992....................................................................33Table 3-5 Regulated California Hazardous Liquid Pipeline Data 1981 through 1990....................................................................35Table 3-6 California Crude Oil Pipelines Under Study 1993 through 1995....................................................................37Table 3-7 Comparison of Various Incident Data Sources..........................................39Table 4-lA Overall Incident Causes Crude Oil Pipelines Under Study......................................................46Table 4-lB Overall Incident Cause Distribution Crude Oil Pipelines Under Study......................................................47Table 4-2A Incident Rates by Year of Study. Regulated California Hazardous Liquid Pipelines......................................49Table 4-2B Incident Rates by Year of Study Crude Oil Pipelines Under Study......................................................51Table 4-3A Incident Rates by Decade of Construction Regulated California Hazardous Liquid Pipelines......................................53.Table 4-3B Incident Rates by Decade of Construction Crude Oil Pipelines Under Study......................................................55Table 4-4A Incident Rates by Normal Operating Temperature Regulated California Hazardous Liquid Pipelines......................................57Table 4-4B Incident Rates by Normal Operating Temperature Crude Oil Pipelines Under Study......................................................58Table 4-5A Incident Rates by Pipe Diameter Regulated California Hazardous Liquid Pipelines......................................60Table 4-5B Incident Rates by Pipe Diameter Crude Oil Pipelines Under Study......................................................62Table 4-6A Incident Rates by Leak Detection System Regulated California Hazardous Liquid Pipelines......................................65Table 4-6B Incident Rates by Leak Detection System Crude Oil Pipelines Under Study......................................................66Table 4-6C Distribution of Pipelines with Leak Detection Systems................................67Table 4-7A Cathodic Protection Systems 5

California State Fire Marshal July 1996STAEflREMAICS14AL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Regulated California Hazardous Liquid Pipelines...........69Table 4-7B Average Cathodic Protection Survey Intervals Regulated California Hazardous Liquid Pipelines...........71Table 4-7C Cathodic Protection System Crude Oil Pipelines Under Study...........................73Table 4-7D Incidents by Cathodic Protection System Inspections Crude Oil Pipelines Under Study...........................75Table 4-8A Incidents by Pipe Specification Regulated California Hazardous Liquid Pipelines...........76Table 4-8B Incident Rates by Pipe Specification Crude Oil Pipeiines Under Study...........................77Table 4-9A Incidents By Pipe Specification Regulated California Hazardous Liquid Pipelines...........79Table 4-9B Incident Rates by Pipe Type Crude Oil Pipelines Under Study...........................82Table 4-1 OA Incident Rates By Normal Operating Pressure Regulated California Hazardous Liquid Pipelines...........84Table 4-lOB Incident Rates by Normal Operating Pressure Crude Oil Pipelines Under Study...........................86Table 4-11 A Incident Rates by Coating Type Regulated California Hazardous Liquid Pipelines...........89Table 4-1IB Incident Rates by External Coating Type Crude Oil Pipelines Under Study...........................91Table 4-1 2A Incidents By Internal Inspections Regulated California Hazardous Liquid Pipelines...........93Table 4-1 2B Incident Rates by Internal Inspections Regulated California Hazardous Liquid Pipelines...........95Table 4-1 3A Incident Rates l.~y Month of Year Regulated California Hazardous Liquid Pipelines...........96Table 4-l3B Incident Rates by Month of Year Crude Oil Pipelines Under Study...........................98Table 4-14 Incidents by Item Which Leaked Regulated California Hazardous Liquid Pipelines...........99Table 4-I 5A Average Hydrostatic Testing Interval During Study Period Regulated California Hazardous Liquid Pipelines..........102Table 4-1 5B Time Since Las~ Hydrostatic Test At Time of Leak Regulated California Hazardous Liquid Pipelines..........103Table 4-1 SC Average Hydrostatic Testing Interval During Study Period Crude Oil Pipelines Under Study..........................104Table 4-16A Spill Size Distribution Regulated California Hazardous Liquid Pipelines..........106Table 4-16B Spill Size Distribution Regulated California Hazardous Liquid Pipelines..........107Table 4-16C Spill Size Distribution Crude Oil Pipelines Under Study..........................109Table 4-16D Spill Size Distribution Crude Oil Pipelines Under Study..........................110 6

~` ~California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STAT~R~ ~tAXSHALTable 4-1 7A Property Damage Distribution Regulated California Hazardous Liquid Pipelines......................................113Table 4-19 Incident Rates by Above versus Below Grade Crude Oil Pipelines Under Study......................................................118Table 6-lA Spill Size Distribution Regulated California Hazardous Liquid versus Crude Oil Pipelines Under Study ....................................................................................135Table 6-lB Property Damage Distribution Regulated California Hazardous Liquid versus Crude Oil Pipelines Under Study ....................................................................................136Exhibit 1 - Acknowledgments 7

r California State Fire Marshal `~ ~ July 1996 ~ MAKSXM. Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Executive Summary The McGrath Lake oil spill in Ventura County stimulated public concern regarding crude oil gathering pipeline safety. This December 22, 1993 incident occurred from a crude oil shipping line. This spill released an estimated 2,200 barrels 42 gallons = 1 barrel of crude oil. The oil surfaced and flowed through a culvert, traveled through 150 feet of woodland and brush, to McGrath Creek, then flowed another 1,200 feet into McGrath Lake. The lake is part of a tidal wetland within a large coastal dune system. One of the results of this incident was the passage of California Assembly Bill 3261 O'Connell as codified in California Government Code Section 51015.05 in 1994. This bill required that the California State Fire Marshal CSFM: * establish and maintain a centralized database containing specific information and data pipeline locations, ownership, age, inspection history', etc. regarding certain crude oil pipelines, * conduct a study of the fitness and safety of these crude oil pipelines, and * investigate incentive options that would encourage pipeline replacement or improvements, including, but not limited to, a review of existing regulatory, permit, and environmental impact report requirements and other existing public policies that could act as barriers to the replacement or improvement of these pipelines. The following pipelines were included in the data base and study: * pipelines for the transportation of crude oil that operate at gravity or at a stress level of 20% or less of the specified minimum yield strength of the pipe; and, pipelines for the transportation of petroleum crude oil in onshore gathering lines located in rural areas. Pipelines meeting this criteria are included in the study and database, whether they were operating or not during the study period; even abandoned, idle, or otherwise out of service pipelines have been included in the study and database. The following pipelines were excluded from the data base and study, in accordance with California Assembly Bill 3261: interstate and intrastate pipelines which are currently regulated by the California State Fire Marshal or the United States Department of Transportation; gathering lines located entirely within the boundary of a California Division of Oil, Gas, and Geothermal Resources DOGGR oil field boundary, or which cross a boundary where two DOGGR oil fields are contiguous and are contained entirely within multiple DOGGR oiljields; 8

California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STATE~1REMA~AL flow lines located entirely within the boundary of a DOGGR designated oil field boundary, or which cross a boundary where two DOGGR oil fields are contiguous and are contained entirely within multiple DOGGR oilfIelds; * natural gas pipelines; * refined petroleum product pipelines; and * abandoned pipelines which have been physically removed. The database and study analyzed California's crude oil gathering pipeline risks, utilizing leak incident data for a three year period, from January 1993 through December 1995. Extensive efforts were taken to gather the most complete database possible. This included the distribution of over 1,200 questionnaires aimed at identif~'ing study participants. An extensive campaign was then conducted to gather as complete a response as possible from the non-responsive operators. This campaign consisted of the following efforts to resolve the nearly 800 non-responsive operators, mostly oil and gas well operators: * initiated thousands of telephone calls to these operators, * mailed and faxed hundreds of additional follow-up letters and questionnaires, solicited industry group assistance i.e. Western States Petroleum Association, Independent Oil Producers Agency, and the California Independent Petroleum * Association, and * worked with DOGGR to pursue additional contacts within the oil and gas well operating companies. Through the end of December 1995, the initial 1,200 potential study participants had been narrowed to a maximum of 740 active pipeline and well operators; the other 460 operators had been deleted primarily because their properties had been sold, or they were duplicate entries. From the remaining 740 operators, a total of 658 responses were received; 82 operators had still not responded to EDM Services, Inc. repeated correspondence. This represented an 89% response through the end of December 1995. At this point, the project schedule was extended. As of April 10, 1996, EDM Services had received a total of 710 responses, from the 739 operators remaining in the database. This represented a 96% response rate. Of these 710 operators, only 15 were identified as owning andlor operating pipelines which met this study criteria. Although the scope of the pipelines meeting the study criteria was originally believed to be quite large, the resulting data set was surprisingly small. Quite simply, there were very few California pipelines which met the study criteria. As a result, few meaningful conclusions could be drawn from this limited data. The data set can be summarized as follows: 9

California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines * Number of Incidents of one barrel or greater - ten 10 * Number of Pipelines - 113 * Total Length of Pipelines - 496 miles Mean Pipe Diameter - 7.5 inches * Mean Operating Temperature - 74 .2°F * Cathodically Protected Pipe - 317 miles 64% of total Bare Pipe - 87 miles 18% of total Median Spill Size - 3 barrels * Average Spill Size - 122 barrels Median Damage - $5,000 $US 1994 * Average Damage - $39,020 $US 1994 * Percentage of Below Grade Pipe - 96.3% 478 miles Overall Incident Rates The overall leak incident rate for incidents of one barrel or more from the crude oil pipelines under study is very similar to the regulated hazardous liquid pipelines - 6.72 versus 6.54 incidents per 1,000 mile years respectively. However, the incident rate for larger spills is generally much less for the smaller, crude oil pipelines in this study. The results for the California crude oil pipelines under study are summarized below:* EventIncident Rate1 barrel or greater spill6.72 incidents per 1,000 mile years10 barrels or greater spill2.02 incidents per 1.000 mile years100 barrels or greater spill1.10 incidents per 1,000 mile years1,000 barrels or greater spill0.69 incidents per 1,000 mile years10,000 barrels or greater spill0.00 incidents per 1,000 mile years$1,000 SUS 1994 damage resulting from spill6.72 incidents per 1,000 mile years$10,000 $US 1994 damage resulting from spill1.34 incidents per 1,000 mile years$100,000 $US 1994 damage resulting from spill1.14 incidents per 1,000 mile years$1,000,000 $US 1994 damage resulting from spill0.00 incidents per 1,000 mile years10

~. ~c¯'~ California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesEventIncident Rateinjury resulting from spill0.00 incidents per 1,000 mile yearsfatality resulting from spill0.00 incident per 1.000 mile years External Corrosion External corrosion is by far the leading cause of incidents, representing 60% of the total. However, with the limited data sample, the cause could not be isolated. The results of the 1993 study regarding the regulated California hazardous liquid pipelines, indicated that pipe operating temperature and age were the two leading factors contributing to increased external corrosion. It can be presumed that this is also the case for the crude oil pipelines under study. However, the data set is too small to perform a conclusive analysis. The overall incident rate for the crude oil pipelines under study is essentially the same as the incident rate for regulated California hazardous liquid pipelines. Although the overall leak incident rates for these groups of pipelines is similar, the likelihood of large spills, and spills resulting in large values of damage, were much lower for the crude oil pipelines in this study. And finally, although the data is limited, there was no evidence to suggest that crude oil spills pose a significant risk to human life. As a result, we recommend that the California Government Code be modified in the following ways: Develop a set of criteria which can be used to identify pipelines which would likely impact unusually sensitive areas in the event of a leak. These criteria might include: likelihood of a spill from a given pipeline to reach a stream or waterway; etc. Distribute this criteria to the owners of the pipelines identified in this study. The operators would then identify those pipelines which would likely impact unusually sensitive areas in the event of a leak. Include the pipelines identified which would likely impact unusually sensitive areas into the existing regulations regarding intrastate hazardous liquid pipelines. * Modify the law to require leak and pipeline inventory reporting for these lines, using the forms developed for this study. The law should also be modified to require leak and pipeline inventory reporting using these same forms for the intrastate and interstate pipelines. This will enable the CSFM to keep the database current. In addition to these statutory recommendations, we recommend the following actions: Continue to invite the operators of these pipelines to attend the regular Safety Seminars and other training programs provided by the CSFM. These programs would also be useful to other local and state public agencies. The database and mapping effort conducted as part of this study should be expanded to include California's intrastate and interstate pipelines. Funding should be appropriated to support this effort. 11

~g. Cahfornia State Fire Marshal`I' July 1996STA1EIREMAZSH.~L Risk Assessment of Califorrtia Low Pressure Crude Oil and Crude Oil Gathering Lines An abbreviated report, covering the items included in Section 4.0 of this study should be prepared every 5 to 10 years. The goals of this study should be to identify incident rate trends, review current regulatory effectiveness, and recommend change. The permitting process for pipeline replacement or upgrade projects should be streamlined to the greatest extend possible. See also Section 7.2 of this report for specific recommendations. Operators should be encouraged to replace older pipelines, when appropriate, to ensure pipeline safety. 12

~California State Fire MarshalJuly 1996 "1kRisk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STATa~~MAXSMAL1.0 Introduction The McGrath Lake oil spill in Ventura County stimulated public concern regarding crude oil gathering pipeline safety. This December 22, 1993 incident occurred from a crude oil shipping line. This spill released an estimated 2,200 barrels 42 gallons = 1 barrel of crude oil. The oil surfaced and flowed through a culvert, traveled through 150 feet of woodland and brush, to McGrath Creek, then flowed another 1,200 feet into McGrath Lake. The lake is part of a tidal wetland within a large coastal dune system. One of the results of this incident was the passage of California Assembly Bill 3261 O'Connell as codified in California Government Code Section 51015.05 in 1994. This bill required that the California State Fire Marshal CSFM: * establish and maintain a centralized database containing specific information and data pipeline locations, ownership, age, inspection history, etc. regarding certain crude oil pipelines, conduct a study of the fitness and safety of these crude oil pipelines, and * investigate incentive options that would encourage pipeline replacement or improvements, including, but not limited to, a review of existing regulatory, permit, and environmental impact report requirements and other existing public policies that could act as barriers to the replacement or improvement of these pipelines. The following pipelines have been included in the data base and study: * pipelines for the transportation of crude oil that operate at gravity or at a stress level of 20% or less of the specified minimum yield strength of the pipe; and, pipelines for the transportation of petroleum crude oil in onshore gathering lines located in rural areas. Pipelines meeting this criteria have been included in the study and database, whether they were operating or not during the study period; even abandoned, idle, or otherwise out of service pipelines have been included in the study and database. The following pipelines were excluded from the data base and study: interstate and intrastate pipelines which are currently regulated by the California State Fire Marshal or the United States Department of Transportation; gathering lines located entirely within the boundary of a California Division of Oil, Gas, and Geothermal Resources DOGGR oil field boundary, or which cross a boundary where two DOGGR oilfields are contiguous and are contained entirely within multiple DOGGR oilfields; 13

California State Fire Marshal July 1996STATEflREMAg~fl~~ Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines flow lines located entirely within the boundary of a DOGGR designated oil field boundary, or which cross a boundary where two DOGGR oilfields are contiguous and are contained entirely within multiple DOGGR 01/fields; * natural gas pipelines; * refined petroleum product pipelines; and * abandoned pipelines which have been physically removed. This report, combined with the completed database, are intended to meet the law's requirements of the CSFM. This report analyzes California's crude oil gathering pipeline risks, utilizing leak incident data from January 1993 through December 1995. The database includes a complete inventory of the pipelines meeting the study criteria, their ownership and location, inspection and maintenance practices, the incidents which occurred from these lines during the study period, and various other data. The study was funded by the U. S. Department of Energy, Bart!esville Project Office, through its Management and Operations contract with BDM Oklahoma, Inc. EDM Services, Inc. conducted this study as a subcontractor to BDM Oklahoma. Brian L. Payne served as project manager and authored this report, except for Section 5.0. Section 5.0 was authored by Deborah Pratt and Jerry R. Simmons of BDM Oklahoma; Jerry Simmons also served as Project Manager. 1.1 Regulatory Authority The California State Fire Marshal CSFM exercises safety regulatory jurisdiction over interstate and intrastate pipelines used for the transportation of hazardous or highly volatile liquid substances within California. In 1983, the Pipeline Safety and Enforcement Program was specifically created to administer this effort. In 1987, the CSFM acquired the regulatory responsibility for interstate lines when an agreement was executed with the United States Department of Transportation USDOT. In doing so, the CSFM became an agent of the USDOT responsible for ensuring that California interstate pipeline operators meet federal pipeline safety standards. Specifically, interstate pipelines under this agreement are subject to the federal Hazardous Liquid Pipeline Safety Act of 1979, as re-authorized in 1992, and federal pipeline regulations. The California State Fire Marshal's responsibility for intrastate lines is covered in the Elder California Pipeline Safety Act of 1981. The CSFM's responsibilities are therefore twofold: First, to enforce federal minimum pipeline safety standards over all regulated interstate hazardous liquid pipelines within California; and 14

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~!~~MAJ~SUAL Secondly, to enforce the above, as well as the Elder California Pipeline Safety Act of 1981 on regulated hazardous liquid intrastate pipelines. The California Division of Oil, Gas, and Geothermal Resources DOGGR has regulatory authority over all oil, gas, and geothermal exploration and production operations in the state. As a part of this authority, the DOGGR has responsibility for regulating flowlines, gathering lines, and other in-field pipelines used to transport crude oil, natural gas, and other fluids. The DOGGR's pipeline jurisdiction ends at the administrative boundary of a field, which is usually the point where ownership of oil or gas is transferred to a pipeline company or oil shipper. As a result, there are crude oil pipelines which are not regulated by any State agency. These pipelines include those which leave DOGGR oil fields and do not meet the pipeline definition of the California Government Code Section 51010.5 the Elder Pipeline Safety Act of 1981. These pipelines are the subject of this study. 1.2 Relative Safety Perspective Before we analyze the risks associated with California's hazardous liquid pipelines, it is important to put the relative safety of pipelines versus other modes of transportation into perspective. The United States Department of Transportation, Research and Special Programs Administration's 1995 National Transportation Statistics - Annual Report provides some useful statistics in this regard. During 1993, there were 43,179 transportation related fatalities in the United States. This data is presented in Table 1-1 by mode of transportation. It should be noted that of the fourteen 1993 pipeline fatalities. Only 0.032% of the total domestic transportation fatalities. All fourteen of them occurred on gas pipelines; there were no fatalities which resulted from incidents on hazardous liquid pipelines. In an attempt to compare the relative safety of each transportation mode, we have estimated the fatality rate per billion ton-miles transported. This was done by first determining the number of 1993 fatalities associated with revenue freight. This was performed for each mode of transportation as follows: Pipelines - All fatalities were included. Rail - All fatalities, including those occurring at grade crossings with vehicular traffic were included. Marine - Recreational boating fatalities were excluded. 15

~g. California State Fire Marshal July 1996SrA ~ Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Air - All general aviation, air taxi, and commuter fatalities were excluded. Since the remaining air carrier data does not differentiate between incidents associated with passenger traffic versus those associated with freight, the resulting number of revenue freight fatalities is unrealistically high. Highway - Only truck fatalities were included. Since truck accidents often result in fatalities to those in automobiles, the resulting truck only fatality figure is unrealistically low. The fatality rate was then determined by dividing the number of fatalities by the number of ton-miles transported. The number of fatalities and resulting fatality rates are presented in Tables 1-2 and 1-3. Despite the inherent data errors, the resulting rates provide a very useful method for determining the relative magnitudes of risk to human life. These results are summarized below, using an arbitrarily assigned risk of I for pipelines. * Pipelines 1 * Marine 5 Rail 51 Highway 429 In other words, rail transportation results in roughly 51 times more fatalities than pipelines for a given number of ton-miles transported. Order of magnitude comparisons between the other modes could be determined similarly. A general understanding of these relative risks is essential for those considering regulatory changes which could increase the cost of hazardous liquid pipeline construction, operation, and/or maintenance. Any increases in the shipping costs associated with such changes would likely result in a portion of the throughput being diverted from pipelines to other transportation modes. Since these other modes generally expose the public to a higher risk than pipelines, any such diversion may actually decrease overall transportation safety. For example, if a costly regulation decreased pipeline accidents by say 10%, but diverted some volume to an alternate, less safe mode of transportation, the new result may be a decrease in overall transportation safety. There are already. signs of this occurring, especially in Southern California. The crude from many of the older production fields which was historically transported by pipeline, has been diverted to truck transportation which has the worst safety record. 16

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesCiC5I050403020 L10 - Table I-I Fatalities by Mode of Transportation1993 National Transportation Statistics40141,349Pipeline904782Rail Marine AirHighway17

~a~4 California State Fire MarshaJ July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesEstimated 1993 Table 1-2Fatalities Associated with Revenue FreightBy Mode of Transportation Table 1-3 Estimated 1993 Fatalities Per Billion Ton-Miles I By Mode of Transportation N/A 9,097 Air Highway10 -8-6-4-2-0-141,349 104PipelineRail MarineN/AAir HighwayI,.~C5U,01210 -8-6-4-2-0-ransported1.230.02IPipelineIRail0.11Marineig

California State Fire MarshalJuly 1996 ~Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~11REM.~SHAL 1.3 Acknowledgments The detailed analyses and data contained in this report and database could not have been gathered and presented without the full support and cooperation from each of the pipeline operators. EDM Services, BDM Oklahoma, the U. S. Department of Energy, the California State Fire Marshal's committee see inside front cover sincerely appreciate each operator's commitment to pipeline safety as evidenced by their time, effort and financial expenditures made to help compile this data. We have attempted to acknowledge the key contacts from each pipeline operating company who worked directly on this project in Exhibit 1; we apologize in advance for any omissions. The Pipeline Assessment Steering Committee members are listed on the inside front cover of this report. 19

California State Fire Marshal July 1996sTATe!IREstAJc~iAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines2.0 Methodology The methodology used to complete this study and compile the database in compliance with California Government Code Section 51015.05 has been outlined in the following subsections. 2.1 Funding and Contracting The California State Fire Marshal, sought United States Department of Energy DOE funding. Funding was granted through the DOE's management and operating contractor for the National Oil Program, BDM Oklahoma. BDM Oklahoma solicited proposals to conduct this study and prepare and compile the database. The proposals were evaluated using three specific assessment criteria: technical approach, management, and cost/price. EDM Services, Inc. was selected as offering the best overall value for this project and was awarded a contract. The resulting contract was executed on May 15, 1995. 2.2 Steering Committee The California State Fire Marshal's Nancy Wolfe to organized the required study and worked with BDM Oklahoma to achieve the objectives of the law. At an organizational meeting, it was decided that a state wide Steering Committee would be required to provide guidance and assist with the study. Industry associations and State and local regulatory agencies were contacted to provide names of individuals who would be invited to participate, as committee members. The first Committee meeting was held in Long Beach, California on November 17, 1994. During the meeting, a project schedule was established, the study parameters were discussed and agreed upon, and the process that BDM Oklahoma would use to select a subcontractor were discussed. A second Steering Committee meeting was held on June 15, 1995, with EDM Services staff, to kick-off the project. At this meeting, the following issues were resolved: The Committee established a definition of the leaks which should be included in this study. The criteria for reporting leaks to the California Office of Emergency Services OES one barrel or more, or any spill onto water, or any spill which could threaten ground water was selected for use. 20

California State Fire Marshal ~July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STATB.n~Et~t.~~HAL The Committee developed a period for collecting leak data. The Committee felt that leak data would not be uniformly available before November 1992, when the OES reporting requirements went into effect. The Steering Committee endorsed a two year study period January 1993, through December 1994 for this study. The Committee decided that all inactive and idle pipelines should be included in the study. Only abandoned lines, which had been physically removed, would be excluded from the study, since they no longer exist. The Committee developed a definition for the pipelines which would - be included in this study. This definition was presented earlier in Section 1.0 of this report. Additional Steering Committee meetings were held on July 19, 1995 and November 13, 1995. The project status was reviewed. The meetings proved to be very helpful. The representatives from government and industry all volunteered to help secure responses from the numerous operators who had not yet responded to the study. In addition to the Steering Committee meetings, EDM Services staff attended and made presentations at the following meetings: - November 9, 1995 Planning Meeting - Sacramento, California November 29, 1995 Legislative Update - Senator O'Connell's Office, Sacramento, California 2.3 Identify Study Participants and Pipelines Approximately 1,200 questionnaires were distributed by EDM Services to potential study participants on June 1, 1995. The mailing list for these notification and identification letters was compiled from the following: the owners and operators of regulated California interstate and intrastate pipelines, the owners and operators of refineries, chemical plants, and terminals located in California, and the owners and operators of all oil and/or gas wells located within the state. 21

`~. ~ California State Fire Marshal July 1996STAThRREMAJ~SJ~AL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The notification letter included the following: * a brief description of the law requiring the study, * a statement that the CSFM intends to use the study results to assess the fitness and safety of the pipelines and develop recommendations to improve, repair or replace proposed pipelines, * notification that EDM Services' personnel would be contacting each operator by mail, telephone, and in some cases visiting selected operators to conduct field audits, * a schematic drawing and description which defined the pipelines under study, a form to be used by each operator to identify a contact who would be responsible for coordinating study activities and to identify whether or not their company owned or operated any pipelines meeting the study criteria, and * notification that EDM Services would be forwarding questionnaires to each operator of pipelines meeting the study criteria, soliciting specific information regarding leak records, pipeline inventory, etc. These initial questionnaires were due for return to EDM Services, Inc. by June 12, 1995. However, through the end of July 1995, only 461 responses had been received, with 43 operators indicating that they owned or operated pipelines which should be included in the study. Having only received responses from about one-third of the operators who received the initial questionnaires, EDM Services initiated an extensive campaign. 2.4 Data Gathering In June 1995, EDM Services developed pipeline inventory and leak data questionnaires. The questionnaires included the pipeline inventory and leak data forms and accompanying instructions. They were used to gather the necessary data. These forms and instructions were reviewed and endorsed by the Steering Committee, CSFM and BDM Oklahoma prior to their distribution and use. On June 30, 1995, EDM Services began distributing copies of the Pipeline Inventory and Leak Data Questionnaires to all operators who had been identified for participation in the study. These documents were then distributed to additional operators as they were identified for inclusion in the study. 22

California State Fire MarshalJuly 1996 -Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~!,~1~~4AL 2.5 Database Development A database, containing the necessary data fields, was established using Microsoft Access database software. The database was structured using three tables. The first contained basic operator data contact name, company name, address, telephone number, pipeline location, year of construction, preventive maintenance activities, leak detection system, etc.. The second contained the pipeline inventory data segment diameter, pipe grade, pipe type, year installed, wall thickness, cathodic protection system, above/below grade, coating type, etc.. The third contained the leak data location, date of leak, probable cause, injury/fatality data, total damage, volume spilled, volume recovered. The pipeline operators forwarded completed Pipeline Inventory and Leak Data Questionnaires to EDM Services. The pipeline inventory and leak data was input into the database as it was received from the pipeline operators. The last of the pipeline and leak data for the pipelines identified for inclusion in the study were received on April 10, 1996. The completed database, Microsoft Access software, and a complete Pentium computer system has been delivered to the CSFM. This system will be used to maintain the database of California's crude oil pipelines. 2.6 Field Audits EDM Services Staff personally visited each operator who owned and/or operated pipelines which met the study criteria. This effort had a very positive impact on the accuracy of the study results. Specifically, a number of operators and pipelines were deleted from the study when it was found that their pipelines did not meet the study criteria. The largest percentage of these pipelines were located entirely within an proposed oil field boundary; as a result, they fell within the DOGGR's jurisdiction and did not meet the study criteria. The second largest category of pipelines deleted from the study were regulated interstate and intrastate pipelines, which were already under the CSFM/USDOT jurisdiction. The audits were also very useful in securing missing and incomplete data from the pipeline operators. Telephone interviews were also conducted to secure missing and incomplete data, resolve inconsistencies, and pursue questionable data. 23

California State Fire Marshal July 1996SThTeRRE MS~W. Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 2.7 Barriers and Incentive Options A questionnaire was prepared to gather information regarding the barriers and incentive options. On January 31, 1996, this questionnaire was distributed to the public agencies having pipeline jurisdiction, interested local agencies, Steering Committee members, interstate pipeline operators, intrastate pipeline operators, and the owners of pipelines meeting this study criteria. The questionnaires requested input on the following: What incentives could be provided to pipeline operators to encourage pipeline replacements or improvements? How could these incentives be implemented? What barriers had been encountered with pipeline replacement or improvement projects? Specifically, what regulatory barriers had been encountered? What specific permit barriers had been encountered? What environmental impact report requirements had been a barrier for pipeline replacement and/or improvement projects? What impact, if any, did these barriers have on the pipeline replacement and/or improvement project e.g. project delay, deferral, elimination, etc.? What were the actual consequences financial, environmental, preventable leaks, public safety, employee safety, etc. of these barriers? Did they impact pipeline safety? What were the potential consequences of these barriers? Case histories of pipeline replacement and/or improvement projects which have been delayed, deferred or canceled because of regulatory, permit or environmental impact barriers were requested. A description of the replacement/improvement project and the barriers encountered was requested. A description of the actual and potential consequences financial, environmental, public safety, employee safety, etc. of the project delay, deferral, or elimination was requested. If pipeline safety was sacrificed, specific details were requested 24

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines ~ regarding how and why it was impacted. Recommendations were requested for removing any of the barriers encountered. The completed questionnaires were forwarded to BDM Oklahoma for review and regulatory analysis. The results of this work are presented later in section 5.0 of this report. 2.8 Potential Data Inconsistencies The importance of an accurate pipeline inventory on the study results can't be overemphasized; the inventory data directly affects the calculated incident rates since it is used in the denominator of the incident rate equation. For example, a ten percent error in the pipeline inventory alone would result in a corresponding ten percent error in the calculated incident rate. 25

California State Fire Marshal July 1996~TATa!~MARSnAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines3.0 Background Pipeline Risk Data A number of sources are available for pipeline incident data. Unfortunately however, few of them include the reliable pipeline inventory necessary to determine meaningful incident rates. In this Section, we have presented results from the following sources: CONCAWE Oil Pipelines Management Group's Special Task Force on Pipeline Spillages OP/STF-1. Performance of Oil Industry Cross Country Pipelines in Western Europe. Statistical Summary of Reported Spillages. 1981 to 1994 annual reports. Line Pipe Research Supervisory Committee of the Pipeline Research Committee of the American Gas Association. An Analysis of Reportable Incidents for Natural Gas Transmission and Gathering Lines 1970 Through June 1984, NG-18 Report Number 158. 1989. Line Pipe Research Supervisory Committee of the Pipeline Research Committee of the American Gas Association. An Analysis of DOT Reportable Incidents for Gas Transmission and Gathering Pipelines for June 1984 Through 1992, NG-18 Report Number 213. 1995. United States Department of Transportation, Research and Special Programs Administration, Office of Pipeline Safety. Annual Report on Pipeline Safety. 1986 through 1992 annual reports. Each of these reports provide pipeline incident data for reportable incidents. However, the criteria for reporting incidents differs for each study. This makes direct comparison of the individual results difficult. On the other hand, it provides a methodology for estimating incident rates for spills meeting various criteria. The following subsections provide a summary of the data contained in each of these reports. The incident rates are shown in units of incidents per 1,000 mile years. This unit provides a means for predicting the number of incidents expected for a given length of line, over a given period of time. For example, if one considered an incident rate of 1.0 incidents per 1,000 mile years; one would expect one incident per year on a 1,000 mile pipeline. If the pipeline was only I mile long, one would expect 1/1,000th of an incident per year, or an incident every 1,000 years. Using these units, frequencies of occurrence can be calculated for any pipeline length and/or time interval. 3.1 CONCAWE - 1981 Through 1994 We have summarized the pipeline results for western European pipelines, as presented in the CONCAWE Performance of Oil Industry Cross Country Pipelines In Western Eurone. Statistical Summary of Reported Spillages, 1981 through 1994 annual reports in Table 3-1. 26

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 3-14STATaPIREMA%SHAZ. European Hazardous Liquid Pipeline Incidents As Reported By CONCAWE 1981 through 19941981 1982 1983 1984 1985Total Pipeline Mileage11,737 11,364 11,240 10,743 10,805Number of Incidents16 10 10 13 7Incident Rate Incidents/bOO Mile Years1.36 0.88 0.89 1.21 0.65Number of lnjunes0 0 0 0 0Injury Rate Injuries/bOO Mile Years0.000 0.000 0.000 0.000 0.000Number of Fatalities0 0 0 0 0Fatality Rate Fatalities/I 000 Mile Years0.000 0.000 0.000 0.000 0.0001985, 1987 ~.` 1988 1989 I 1990Total Pipeline MileageNumber of IncidentsIncident Rate Incidents/i 000 Mile YearsNumber of InjuriesInjury Rate Injuries/i 000 Mile YearsNumber of FatalitiesFatality Rate Fatalities/I 000 Mile Years10,805121.11010,80580.74010,99211 * 1.0001991I.1992.:19931994pTotal Mileage IncidentsIncidents/bOO Mite Years 13,049 . 14 1.07:13,359 70.5213,422 100.75*19,138 ii0.57171,220 146 0.85of Injuries00001000 Mile Years0.0000.0000.0000.0000.006Fatalities0*0'00*3000 Mile Years0.0000.0000.0000.0000.01811,737131.11112,02440.3300.00000.00000.000 0.0000.00000.08530.00000.000 0.256 . 0.000Reportable incidents include:1. All leaks greater than one cubic meter 264 gallons or approximately 6 barrels.2. All leaks under one cubic meter which resulted in noteworthy environmental impact.27

California State Fire Marshal~ ~ July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The criteria for including hazardous liquid pipeline incidents in these reports are as follows: all spills greater than one cubic meter approximately 264 gallons or 6 barrels and spills less than one cubic meter, if the spill had a noteworthy impact on the environment. The reader should note that only onshore pipelines were included in these data. Also, beginning in 1994, non-commercially owned pipelines began to be included in the database. It is interesting to note that this reporting criteria does not include any consideration for incidents which cause injuries and/or fatalities. As a result, the injury and fatality incident rates derived from this data may be low. Also, the overall incident rates for these relatively large spills is comparatively low, as shown below: Incident Rate 0.85 incidents per 1,000 mile years Injury Rate 0.006 injuries per 1,000 mile years Fatality Rate 0.0 18 fatalities per 1,000 mile years 3.2 U.S. Natural Gas Transmission and Gathering Lines 1970 Through June 1984 Table 3-2 presents the reportable domestic natural gas transmission and gathering line incidents from 1970 through June 1984. Although this data is for natural gas lines, instead of crude oil lines which are the subject of this study, the data is worth noting for comparison. These natural gas transmission lines are of similar construction to the steel pipelines included in this study. The criteria for leaks to be reported to the USDOT for inclusion in this data are as follows: resulted in a death or injury requiring hospitalization, required the removal from service of any segment of a transmission pipeline, resulted in gas ignition, caused an estimated damage to the property owner, or of others, or both, of $5,000 or more, involved a leak requiring immediate repair, involved a test failure that occurred while testing either with gas or another test medium, or in the judgement of the operator, was significant even though it did not meet any of the above criteria. 28

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 3-2 U. S. Natural Gas Transmission and Gathering Lines Reportable Incidents 1970 through June 1984sTA1.k_~:~HM.1970 1971 1972 1973 1974 1975Total Pipeline Mileage284,196 285,482 285,575 285,241 293,885 267,079Number of Incidents343 409 409 471 458 366Incident Rate Incidents/i 000 Mile Years1.21 1.43 1.43 1.65 1.56 1.37Numberof Injuries24 24 37 19 21 21Injury Rate Injuries/i 000 Mile Years0.084 0.084 0.130 0.067 0.071 0.079Number of Fatalities 1 3 6 2 4 7Fatality Rate Fatalities/i 000 Mile Years 0.004 0.011 0.021 0.007 0.014 0.026Total Pipeline Mileage*thImh~r of ~~nMNumber of Injuries1975 1Q77 I 1-075 1070 I iQRflIn.i.4~n~ ~ ~I ~ 000 Mile Years0.92 1.57 1.46 1.55Intur.j Rate I `~`-I1fl0fl Mile Years~ 0151 00Th 0099 0~09.277.555283,373303,355311,09825444544448242223096-----Iu.,co~ -.___ - --------- - ----- ----- --- - -Number of Fatalities7812Fatality Rate F~+~lih1~I1,000 Mile Years 0.025 0.028 0.003 0.039388,8573250.84160 0410.003Total Pipeline MileageSi. ~I~Ul I lk~CI ~. IIIL.IUCI ItaIncident Rate Inr'ttlont'/l .000 Mile YearsNumber of InjuriesInjury Rate Iniuries/i .000 Mile YearsNumber of F~1i1~~-----. 1-saic t, ata,,L,caq - - - - - - - -- - - - - -- -~R0:~on47~.204-0.971.141.371.2964125ii0.0150.120 0.072 0.07061022 j 1951 1052 1QR~ I 1054 I~ Total400,243 342,645 346,355 157,921 4,512,860S 5571.304350.09672Fatality ~ ` `~`~`1000 Mile Vearsi0015 0029 0005 0 01~ 0015Notes:1. 36 of the total 72 fatalities were to employees of the operating company.2. 161 of the total 274 injuries were to employees of the operating company.3. The 1984 mileage figure shown is one-half the actual mileage to account for only one-half year of data.Reportable incidents include:1. Resulted in a death or injury requiring hospitalization.2. Required the removal from service of any segment of a transmission pipeline.3. Resulted in gas ignition.4. Caused an estimated damaged to the property of the operator, or of others, or both, of $5,000 or more.5. Involved a leak requiring immediate repair.-6. Involved a test failure that occurred while testing either with gas or another test medium.7. Or, in the judgement of the operator, was significant even though it did not meet any of the above cntena.29

California State Fire Marshal & July 1996STA~FuEMA*S~Lkt Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The incident rates for reported leaks meeting this criteria are summarized below: * Incident Rate 1.30 incidents per 1,000 mile years Injury Rate 0.096 injuries per 1,000 mile years * Fatality Rate 0.0 16 fatalities per 1,000 mile years 3.3 U.S. Natural Gas Transmission and Gathering Lines June 1984 through 1992 Table 3-3 presents the reportable domestic natural gas transmission and gathering line incidents from June 1984 through 1992. It is important to note that in June 1984, the USDOT changed the criteria for reporting leaks. The most significant change was that in general, leaks causing less than $50,000 property damage, did not have to be reported. Since this value is significantly greater than the $5,000 criteria for the earlier study period, we see a significant decrease in the resulting reportable incident rate. Although impossible to verif~' using this data, we also believe that the actual frequency of incidents decreased during this period as a result of one-call system implementation, among other things. The criteria for leaks to be reported to the USDOT from June 1984 through 1992 were as follows: Events which involved a release of gas from a pipeline, or of LNG or gas from an LNG facility, which caused: a a fatality, or personal injury necessitating inpatient hospitalization; or b estimated property damage, including costs of gas lost by the operator, or others, or both, of $50,000 or more. An event which resulted in an emergency shut-down of an LNG facility. An event that was significant, in the judgement of the operator, even though it did not meet the criteria above. The incident rates for reported leaks meeting this criteria from June 1984 through 1992 are summarized below: Incident Rate 0.26 incidents per 1,000 mile years Injury Rate 0.06 1 injuries per 1,000 mile years Fatality Rate 0.018 fatalities per 1,000 mile years As demonstrated by the approximately 80% reduction in the incident rate over the earlier period, we see that the change in reporting criteria, among other things, had a major influence on the results. However, it is interesting to note that the injury and fatality rates remained nearly unchanged from the earlier period. 30

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines ~ Table 3-3 Onshore U. S. Natural Gas Transmission and Gathering Lines Reportable Incidents June 1984 through 199219841 1985 .1986 1987 1988Total Pipeline Mileage157,921 324,426 340,202 290,176 310,079Numberof Incidents82 115 77 59 80Incident Rate Incidents/i .000 Mile Years0.52 0.35 0.23 0.20 0.26Number of Injuries32 12 20 15 13lnjuryRatelnjuries/1,000MileYears0.203 0.037 0.059 0.052 0.042Number of Fatalities7 6 6 0 3Fatality Rate Fatalities/I ;000 Mile Years0.044 0.018 0.018 0.000 0.0101989 . I 1990~ 1. .1991 I 1992 j TotalTotal Pipeline Mileage313,751 294,504 315,290 327,484 2,673,833Number of Incidents83 72 65 52 685Incident Rate Incidents/I ,000 Mile Years0.26 0.24 0.21 0.16 0.26Numberoflnjuiies28 17 12 15 164Injury Rate Injuries/i .000 Mile Years0.089 0.058 0.038 0.046 0.061Number of Fatalities22 0 0 3 47Fatality Rate Fatalities/i ,000 Mile Years0.070 0.000 0.000 0.009 0.018Notes:1. The 1984 mileage figure shown is one-half the actual mileage to account for only one-half year of data.Reportable incidents include: Events which involve a release of gas from a pipeline, or of LNG or gas from an LNG facility, which cause a a fatality, or personal injury necessitating inpatient hospitalization; or b estimated property damage, including costs of gas lost by the operator or others, or both, of $50,000 or more. An event which results in an emergency shut-down of an LNG facility. An event that is significant, in the judgement of the operator, even though it did not meet the criteria of 1 or 2 above. 31

California State Fire Marshal July 1996TAT RaE MA~SHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 3.4 U.S. Hazardous Liquid Pipeline Accidents - 1986 through 1992 As noted earlier, a reliable pipeline inventory is necessary to determine precise incident rates. The degree of accuracy of the domestic hazardous liquid pipeline inventory is questionable. For example, the total reported pipeline length remained constant for each year examined. However, we are aware of new line construction and line abandonments during this period. As a result, we believe that the incident rates derived using the reported pipeline lengths are approximations only; they should not be taken as absolute. Table 3-4 presents the reportable domestic hazardous liquid pipeline incidents from 1986 through 1992. The criteria for incidents to be reported to the USDOT for inclusion in this data were as follows: * explosion or fire not intentionally set by the operator, * loss of more than 50 barrels of liquid or carbon dioxide, * escape to the atmosphere of more than five barrels per day of highly volatile liquid, * death of any person, * bodily harm to any person resulting in loss of consciousness, necessity to carry the person from the scene, or disability which prevents the discharge of normal duties or the pursuit of normal activities beyond the day of the accident, and/or * estimated property damage to the property of the operator, or others, or both, exceeding $5,000. The approximate incident rates for reported leaks meeting this criteria are summarized below: * Incident Rate 1.31 incidents per 1,000 mile years * Injury Rate 0.149 injuries per 1,000 mile years * Fatality Rate 0.017 fatalities per 1,000 mile years It's interesting to note that these results are essentially the same as those for reportable U.S. natural gas lines from 1970 through June 1984, which had a similar $5,000 property damage reporting requirement. 32

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines ~ Table 3-4 U. S. Hazardous Liquid Pipeline Accidents Reportable Incidents 1986 through 19921986 .1987 I 1988 1989Total Pipeline Mileage 1150,000 155,000 155,000 155,000Numberof Incidents203 237 196 161Inadent Rate Incidents/i .000 Mile Years1.35 1.53 1.26 1.04Number of Injunes32 20 19 38Injury Rate lnjuries/1,000 Mile Years0.213 0.129 0.123 0.245Number of Fatalities3 3 2 2Fatality Rate Fatalities/i .000 Mile Years0.020 0.019 0.013 0.013... 1990. ~.. . .1 :..: . 1991 .. 1 . .1992 TotalTotal.Pipeline Mileage151,000 152,300 152,300 1,070,600Numberof Incidents ..177 210 223 1,407Incident Rate lncidentsli,000Mile Years~1.17 1.38 1.46 . 1.31 - . Number of Injuries . . .7 5 38 159Injury Ratelnjuries/i,000 Mile Years0.046 0.033 0.250 0.149Number of Fatalities :3 . 0 . 5 18Fatality Rate FaialitiesIl,000 Mile Years0.020 0.000 0.033 0.017Notes:1. The mileage figures are approximate as reported by the U.S. Department of Transportation's, Annual Report on Pipeline Safety, published for each subject year.After October 21, 1985, reportable incidents include: Explosion or fire not intentionally set by the operator. Loss of more than 50 barrels of liquid or carbon dioxide. Escape to the atmosphere of more than five barrels per day of highly volatile liquid. Death of any person. Bodily harm to any person resulting in loss of consciousness, necessity to carry the person from the scene, or disability which prevents the discharge of normal duties or the pursuit of normal activities beyond the day of the accident. Estimated property damage to the property of the operator or others, or both exceeds $5,000. 33

STATE FIRE MARSHAl. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 3.5 Data Summary of Regulated California Hazardous Liquid Pipelines 1981 through 1990 This study included all regulated California interstate and intrastate hazardous liquid pipelines. The systems included in this study had complete leak records. All leaks, regardless of size, extent ofproper!y damage, or extent of injury were included in the study. Also, a complete audit of the pipeline inventory and leak data was conducted. As a result, the incident rates were much higher than presented in earlier studies, which only included reported leaks fitting a relatively narrow criteria. A summary of these results is included in Table 3-5. The incident rates for all leaks, as well as those meeting the noted criteria, which occurred during the ten year study period are summarized below. All financial data has been converted to $US 1994; the incident rates corresponding to various dollar amounts has been estimated using the available data.Incident Rate All LeaksIncident Rate All Leaks -Crude Oil Pipelines OnlyIncident Rate > $1,000Incident Rate > $10,000Incident Rate > $100,000Injury Rate any severityFatality Rate7.08 incidents per 1,000 mile years9.89 incidents per 1,000 mile years5.80 incidents per 1,000 mile years3.64 incidents per 1,000 mile years1.36 incidents per 1,000 mile years0.685 injuries per 1,000 mile years0.042 fatalities per 1,000 mileyears 3.6 Data Summary of California Crude Oil Pipelines Under Study 1993 through 1995 This study included all California crude oil liquid pipelines not previously regulated by any state agency. The systems included in this study had complete leak records. Leak incidents of one barrel or more, or any spill onto water, or any spill which could threaten ground water were included in this study. The leak data was also completely audited. The incident rates were very similar to the results for regulated California hazardous liquid pipelines. A summary of these results is included in Table 3-6. The incident rates for the leaks which occurred during the three year study period are summarized below. All financial data is shown in constant $US 1994.Incident Rate Leaks> I bblIncident Rate > $1,000Incident Rate > $10,000Incident Rate > $100,000Injury Rate any severityFatality Rate6.72 incidents per 1,000 mile years6.72 incidents per 1,000 mile years1.34 incidents per 1,000 mile years1.14 incidents per 1,000 mile years0.00 injuries per 1,000 mile years0.00 fatalities per 1,000 mile years34

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 3-5 Regulated California Hazardous Liquid Pipeline Data - All Leaks 1981 through 19901981 1982 1983 1984 I 19851986Total Pipeline Mileage 6,482 6,658 6,675 6,835 7,0057,501Number of Incidents 53 83 53 30 4546lncidentRatelncidents/1,000MileYears 8.18 12.47 7.94 4.39 6.426.13Number of Injuries 0 1 2 0 015Injury Rate lriiuries/i .000 Mile Years 0.000 0.150 0.300 0.000 0.0002.000Number of Fatalities 0 0 0 0 a:1Fatality Rate Fatalities/i .000 Mile Years 0.000 0.000 0.000 0.000 0.0000.1331987 1988 1989 I 1990I..__TotalTotalPipetineMileage7,587 7,600 7,609 7,61071,563Number of Incidents60 52 42 43507Incident Rate Incidents/i .000 Mile Years7.91 6.84 5.52 5.657.08Number of Injuries0 0 31 049Iniury Rate Injuries/I 000 Mile Years0.000 0.000 4.074 0.0000.685Number of Fatalities0: 0 2 0:3Fatality Rate Fatalities/i .000 Mile Years0.000 0.000 0.263 0.0000.042Note: The above table includes aD leaks, regardless of size or severity. Regulated California Hazardous Liquid Pipeline DataLeaks Greater Than 5 Barrels, or Greater Than $5,000 Damage 1981 through 19901981 1982 1983 1984 19851986Total Pipeline Mileage6,482 6,658 6,675 6,835 7.0057,501Number of Incidents52 73 44 30 4140Incident Rate Incidents/i .000 Mile Years8.02 10.96 6.59 4.39 5.855.33Number of Injuries0 1 2. 0 0.15 * Injury Rate Injuries/i .000 Mile Years0.000 0.150 0.300 0.000 0.0002.000Number of Fatalities0 0 0 0 01Fatality Rate Fatalities/i .000 Mile Years 0.000 0.000 0.000 0.000 0.0000.1331987 I 1988 I 1989 1990TotalTotal Pipeline Mileage7,587 7,600 7,609 7,61071,563Number of Incidents48 42 35 36441 * Incident Rate Incidents/i .000 Mile Years6.33 5.53 4.60 4.736.16Number of Injuries0 0 31 049Iniury Rate Injuries/i .000 Mile Years0.000 0.000 4.074 0.0000.685Number of Fatalities 0 0 2 0.3Fatality Rate Fatalities/I .000 Mile Years 0.000 0.000 0.263 0.0000.042Note: The above table also includes all leaks which resulted in any injury, regardless of severity, and all leaks resulting in fatalities.35

STATE FIRE MARSHAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Regulated California Hazardous Liquid Pipeline Data Leaks Greater Than $50,000 Damage 1981 through 19901981 1982 1983 1984 1985 1986Total Pipeline Mileage 6,482 6.658 6,675 6.835 7,005 7,501Number of Incidents 39 56 33 20 31 27Incident Rate Incidents/i 000 Mile Years 6.02 8.41 4.94 2.93 4.43 3.60Number of Injuries 0 1 2 0 0 15Iniury Rate Injuries/I .000 Mile Years 0.000 0.150 0.300 0.000 0.000 2.001Number of Fatalities 0 0 0 0 0Fatality Rate Fatalities/i .000 Mile Years 0.000 0.000 0.000 0.000 0.000 0.1331987 1988 1989 i990_....L. Total .Total Pipeline Mileage7,587 7,600 7,609 7,610~ 71 .563Number of Incidents34 30 21 26 317incident Rate Incidents/i .000 Mile Years4.48 3.95 2.76 3.42 4.43Numberof Injuries0. 0 31 0 49Injury Rate Injuries/i .000 Mile Years0.000 0.000 4.074 0.000 0.685Number of Fatalities0 0 2 0 3Fatality Rate Fatalities/I .000 Mile Years 0.000 0.000 0.263 0.000 0.042Regulated California Hazardous Liquid Pipeline Data Leaks Greater Than $500,000 Damage 1981 through 19901981i 1982 1983 1984 198~._....L I986__TotalPipelineMileage 6,482 6,658 6,675 6,835 7,005 7,501 ,Number of Incidents 36 50 30 19 ` 28 21Incident Rate Incidents/i .000 Mile Years 5.55 7.51 4.49 2.78 4.00 2.80Number of Injuries 0 1 2 0. 0. 15 -Injury Rate Injuries/i .000 Mile Years 0.000 0.150' 0.300 0.000 0.000 2.000Number of Fatalities 0 0 0 0 0.Fatality Rate Fatalities/I .000 Mile Years 0.000 0.000 0.000 0.000 0.000 0.1331987 1988 1989 1990 TotalTotal Pipeline Mileage 7,587 7,600 7,609 7,610 71,563Number of Incidents 31 24 18 24 281Incident Rate Incidents/i .000 Mile Years 4.09 3.16 2.37 3.15 3.93Number of Injuries 0 0 31 0 49Injury Rate lnjunesli.000 Mile Years 0.000 0.000 4.074 0.000 0.685Number of Fatalities 0 0 2 0' 3Fatality Rate Fatalities/i .000 Mile Years 0.000 0.000 0.263 0.000 0.042Note: The above tables also include all leaks WrUCI'~ resulted in any injury, regardless of severity, and all leaks resulting in fatalities.

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesSTAT~!~~~M. Table 3-6 Data for California Crude Oil Pipelines Under Study - All Leaks 1993 through 1995Note: The above table includes all leaks >1 barrel.1993 :~ 1994 1995 To~l~ot~pip~ine Mi~age `1 494 496 496 1486~ 1 4 5 10.1 ~2.02 8.06 10.08 6.72Numberof.lnju~e~ : r0 0 0 00.000 0.000 0.000 0Number of Fatalities0 0 0 0~FatalityR‡teFat·IitiËs/1,OOOMile'Years 0.000 0.000 0.000 037

California State Fire Marshal July 1996srATEnR1Md~cSnAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Although the overall incident rates for this study were very similar to those recorded in the earlier regulated California hazardous liquid pipeline study 6.72 versus 7.08 incidents per 1,000 mile years, it's interesting to note that the incident rates for spills resulting in various amounts of damage were significantly lower, as indicated below.DescriptionCalifornia Crude OilPipelines Under StudyRegulated CaliforniaHazardous Liquid Pipelines Incident Rate> $1,000 Damage $US 19946.72 Incidents per1,000 mile years5.80 Incidents per1,000 mile years Incident Rate> $10,000 Damage SUS 19941.34 Incidents per1,000 mile years3.64 Incidents per1,000 mile years Incident Rate> $100,000 Damage SUS 19941.14 Incidents per1,000 mile years1.36 Incidents per1,000 mile years Incident Rate> $ 1,000.000 Damage $US 19940.00 Incidents per1,000 mile years0.28 Incidents per1,000 mile years This parameter will be reviewed in more detail later in this report. However, this result is reasonable, since the crude oil pipelines under study are generally much smaller in diameter and length, are primarily located in rural areas, and do not transport refined petroleum products. 3.7 Comparison of Various Incident Data Sources Table 3-7 demonstrates the differences that various reporting criteria have on the resulting incident rates. It should be noted that the California incident rates, which appear to be much higher, are the only data which have been completely audited. These data do not necessarily indicate that California's pipeline network presents a higher risk than those in other areas. Unfortunately however, we could not find audited data from other areas, with complete leak records, for comparison. One of the benefits of having data available which met various reporting standards was that incident rates could be established for a variety of criteria. For example, the regulated California hazardous liquid data could be used to establish incident rates for all leaks and injuries. Data from the other studies could be used to establish incident rates for their specific reporting criteria. These differences are summarized in the following subsection. 38

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 3-7 Comparison of Various Incident Data Sources Incidents Per 1,000 Mile YearsETA FIRE MA~.HAI. * ln~dent lnju~ Fa~l~ * Rate * Rate :~te:CONCAWE -1981 to 19940.850 0.010 0.030 * U;S. Natural Gas -1970toi984.:~:1.300 0.096 0.0161:. u.S Nat alGa J1984tci1992 0.260 0.061 0.018 LJ.S.t1athdÙusLiqt~kt986.to1992..~I 1.310 0.149 0.017RÈgulatedCartf~Ha Liq. ill Ieak~-i98tto 19907.080 0.685 0.042~artfomiaCrudeOWPipes Under St¸dy-i993tc 1995j 6.720 0.000 0.000~Calfornia leak&> 5bbl,or>$5~O00-1981 to 1990 3.360 0.000 0.000 ~:~lifom~ L ks $50,000 -1981 tO 1990 0.670 0.000 0.000 2~ 0 CONCAWE `81-94 US Gas Á84-92 Reg CA Haz Liq p81-90 US Gas 70-84 US Liquid `86-92 CA Crude Study `93-95Note: The California regulated hazardous liquid pipeline data includes all leaks and injuries, regardless of severity.Further, the California data was completely audited. The resulting higher California incident rates do not necessarilyindicate that California crude oil and/or regulated hazardous liquid pipelines pose a higher risk than those included inother studies. The reader should consult the report text for a more complete discussion.Incident Rate Comparison Incidents Per 1,000 Mile Years1064-I39

STATh~IRE MA~AL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathermg Lines 1.0 0.8 0.6 0.4 0.2 0.0 0.10 0.08 0.06~ 0.04 0.02 Injury Rate Comparison Injuries Per 1,000 Mile Years - ..,- ____z CONCAWE `81-94 US Gas `84-92 Reg CA Haz Liq `81-90 US Gas `70-84 US Liquid `86-92 CA Crude Study `93-95 Fatality Rate Comparison Fatalities Per 1,000 Mile Years CONCAWE `81-94 US Gas `84-92 Reg CA Haz Liq `81.90 US Gas `70-84 US Liquid `86-92 CA Crude Study `93-95Note: The California regulated hazardous liquid pipeline data includes all leaks and injuries, regardless of severity.Further, the California data was completely audited. The resulting higher California incident rates do not necessarilyindicate that California crude oil and/or regulated hazardous liquid pipelines pose a higher risk than those included inother studies. The reader should consult the report text for a more complete discussion.0.00~-4n

.~q.California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STAT~R~~ L 3.8 Uncorrected Pipeline Risks Using the data developed in the prior subsections, one can estimate the incident rates for various pipeline events as follows:EventIncident Rateany size leak from regulated hazardous liquid pipeline7.1 incidents per 1,000 mile yearsone barrel or larger leak from crude oil pipeline under study6.72 incidents per 1,000 mile yearsproperty damage greater than $1,000 $US 19946.7 incidents per 1,000 mile yearsproperty damage greater than $10,000 $US 19941.3 to 3.6 incidents per 1,000 mile yearsproperty damage greater than $100,000 SUS 19941.1 to 1.4 incidents per 1,000 mile yearsproperty damage greater than $1,000,000 SUS 19940.00 to 0.28 incidents per 1,000 mile yearsany injury0.0 to 0.70 injuries per 1,000 mile yearsinjury requiring hospitalization0.0 to 0.10 injuries per 1,000 mile yearsfatality0.0 to 0.04 fatalities per 1,000 mile years These values may be useful when evaluating the risks associated with proposed pipeline projects. However, as noted by the wide range of values presented, the user should use judgement in selecting the appropriate values for a particular project. Consideration should be given to the type of pipeline under investigation, the contents being transported, pipe age, type of coating, operating temperature, and other parameters. The data presented in Section 4.0 of this report, and the 1993 California Hazardous Liquid Pipeline Risk Assessment will aid the reader in making such assessments. 41

"~ a~ California State Fire Marshal July 1996 snr~r~*icsa*j. Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 4.0 General Risk Levels Before reviewing the specific study results, it is helpfiul to review a profile of the crude oil pipelines included in this study. To reiterate the information presented earlier in Section 1.0, the following pipelines have been included in this study and database: pipelines for the transportation of crude oil that operate at gravity or at a stress level of 20% or less of the specified minimum yield strength of the pipe; and, pipelines for the transportation of petroleum crude oil in onshore gathering lines located in rural areas. Pipelines meeting this criteria have been included in the study and database, whether they were operating or not during the study period; even abandoned, idle, or otherwise out of service pipelines have been included in the study and database. The following pipelines were excluded from the data base and study: interstate and intrastate pipelines which are currently regulated by the CSFM or USDOT; gathering lines located entirely within the boundary of DOGGR oil field boundary, or which cross a boundary where two DOGGR oil fields are contiguous and are contained entirely within multiple DOGGR oil fields; flow lines located entirely within the boundary of a DOGGR designated oil field boundary, or which cross a boundary where two DOGGR oil fields are contiguous and are contained entirely within multiple DOGGR oil fields; natural gas pipelines; refined petroleum product pipelines; and abandoned pipelines which have been physically removed. It's also important to understand the leak incidents which have been included this study. As noted earlier, the criteria for defining these leaks was established by the Steering Committee. The criteria for reporting leaks to the California Office of Emergency Services OES one barrel or more, or any spill onto water, or any spill which could threaten ground water was selected for use. Unfortunately, the OES spill database could not be used for this study, since it does not contain sufficient pipeline and leak details to facilitate any specific analyses. The study period was established as a three year period from January 1993 through December 1995. 42

CI~California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STAT~FIREMA~}LAL Although over 1,200 questionnaires were initially distributed to potential study participants, the actual number of leaks and the length of crude oil pipelines included in this study is relatively small; there are simply very few miles of pipeline which met the study criteria. This data set only included ten 10 leaks of one barrel or greater, which occurred during the three year study period, from only 496 miles of pipelines. This data sample is simply too small to draw many meaningful conclusions. Despite the instructions requesting that only leaks of one barrel or greater be reported except for those meeting other criteria we received ten leak reports for spills of less than one barrel. Since this data was not uniformly available or reported for all of the operators, these incidents of less than one barrel were not included in the study. It's worth noting that the total damage from these leaks, which were excluded from the study, was nominal, averaging $3,460 per incident. For comparison purposes, we have also presented data for regulated California hazardous liquid pipelines, as reported in the 1993 California Hazardous Liquid Pipeline Risk Assessment. Throughout this section, comparisons have been made between California's crude oil pipelines under study and the regulated California hazardous liquid pipelines, for reference. Profiles of these pipeline data sets are summarized below:California Crude Oil`Pipelines Under Study ..Regulated CaliforniaHazardous Liquid. Pipelinesof Pipelines 496 Milesin Study7,8O0 MilesIncident Data 3 Years 1993-1995 10 Years1981-1990Internal Inspection 28 MilesIncluded in Study 5.6%4,495 Miles 57.6%Pipelines or Line 113 Pipelines in Study552 Pipelinesof Each Pipeline 4.39 Miles14.1 MilesOriginal Pipe 19531957of Pipe 7.5 Inches12.3 Inchesof Internal 15.1 Inches14.3 Inches External Corrosionof Incidents 60% of all Leak Incidents External Corrosion59% of all Leak Incidents 1.3 Miles of Bare andUncoated Pipe 149 Miles of Unknown 0.3% bare and 30% unknown 530 Miles6.8% of Total 317 MilesProtected Pipe 64% of Total 6,976 Miles99.4% of Total43

California State Fire Marshal July 1996STATflITREMASHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesDescriptionCalifornia Crude OilPipelines Under StudyRegulated CaliforniaHazardous Liquid PipelinesMean Normal Operating Temperature74.2°F97.9°FNumber of Leaks During Study PeriodlOLeaks5l4LeaksAverage Spill Size122.1 Barrels408 BarrelsMedian Spill Size3 Barrels5 BarrelsAverage Damage Per Incident Uninflated$39,000 SUS 1994$211,000 $US 1994Median Damage Per Incident$5,000 $US 1994$10,710 $US 1994Average Age Of Leak Pipe39.9 Years40.8 YearsAverage Diameter of Leak Pipe7.5 Inches10.2 Inches Mean Normal Operating , Temperature of Leak Pipe64.5°F109.7°FInjuries During Study Period049Fatalities During Study Period03 In the table above, the terms mean and average were used to differentiate between the methods used to calculate the values. Average values were determined by simple division. For example, the average spill size was determined by dividing the sum of each individual spill volume by the total number of spills. Mean values, on the other hand, were determined by weighting the individual parameters by pipe length and the number of years of service during the study period. For instance, the mean normal operating temperature was determined as follows: Tme~ = E {T1L~Y~ + TI+ILI+IY~+l + ...} ˜ E {L1Y1 + Ll+IYI+1 + where: Tme~ = mean normal operating temperature T1 = normal operating temperature for line segment~ L, = length of line segment~ = number of years of line segment1 operation during study period We believe that this weighting method provides a much more meaningfi.il representation of mean values for many parameters than simple division. It has been used where appropriate to determine the values shown in many of the tables presented in this report. 44

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 4.1 Overall Incident Causes The overall incident rate for all pipelines included in this study was 6.72 incidents one barrel or greater per 1,000 mile years. Table 4-IA presents the detailed data. As indicated, the leading cause of leak incidents of California's crude oil pipelines under study from January 1993 through December 1995 was external corrosion, which caused 60 percent of all leaks. The second leading factor was internal corrosion, which caused 20% of all leaks. The volumes spilled as a result of external corrosion were nominal in size, relative to the spill size resulting from other causes 3 barrel average for external corrosion, versus 300 barrel average for other causes. The remaining 20% of the leaks were caused by third party damage, distributed equally 10% each between the following: third party damage due to construction, and third party damage due to farm equipment. The incident cause distribution for California's crude oil pipelines under study and regulated California hazardous liquid pipelines are compared numerically below, and graphically in Table 4-lB. CaliforniaCrude OilIncident. Pipelines Under StudyRegulated CaliforniaHazardous Liquid PipelinesCorrosion 60%59%Corrosion 20%3%Party 20%20%Malfunction 0%5%Failure 0%4%Error 0%2%0%10% As shown, external corrosion caused the majority of the leak incidents in both data sets. The issues regarding this cause of leaks will be explored in more detail in many of the following subsections of this report. 45

`V STATER~ MARSHAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-IA Overall Incident Causes California Crude Oil Pipelines Under Study Incident Rate Comparison Incidents Per 1000 Mile YearsNo. of Incidents Rate6 4 03Percentage260% 1 06720% 1 ~6710%10% io 672 100%1,4871953742 -75122 1 Incident Cause Distribution California Crude Oil Pipelines Under Study60.0% External Corrosion20.0% Internal Corrosio~ ~-"~` 10.0% 3rd Party - Farm Equipmen~i-~' 10.0% 3rd Party - ConstructionI46

California State Fire MarshalJuly 1996Risk Msessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines a, C Co C C- 0 a- Table 4-lB Overall Incident Cause DistributionRegulated California Hazardous Liquid Pipelines versus California Crude Oil Pipelines Under Study- External Corrosion Internal Corrosion~ 3rd Party - AU~ Operator Error Equipment Malfunction~ OtherReg. Calif. Haz. Liq. Cal. Crude Oil Under Study47

California State Fire Marshal July 1996STAT~FTREMAj~sH~L Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Internal corrosion caused a much larger percentage of the pipeline incidents under study 20% versus only 3% for the regulated California hazardous liquid pipeline incidents. This is not surprising, since many of these pipelines are crude oil gathering lines. As a result, one would expect that they carry a higher percentage of water and other impurities which would tend to increase the internal corrosion rate. In fact, many of these lines 330 miles, 67% transport crude oil with water cuts between 1% and 3%; 19 miles 4% transport crude oil with water cuts greater than 3%. This is in contrast to nearly all of the regulated transmission lines, which typically transport crude oil with less than 1% water. The remaining 29% of the pipelines under study did not report this parameter. Third party damage caused 20% of the pipeline incidents in this study. This is the same distribution as the regulated California hazardous liquid pipelines. 4.2 Incident Rates By Study Year Regulated California Hazardous Liquid Pipelines For the regulated California hazardous liquid pipelines, varying leak incident rates were observed during the ten year study period. Table 4-2A shows the incident rate break-down for each year during the ten year survey period by cause. The results demonstrate a slight decline over the ten year period: during the first five years the average incident rate was 8.5; during the latter half the average incident rate was 6.9 leaks per 1,000 mile years. An ordinary least squares line of best fit was determined to evaluate the statistical relevance of this overall leak data by year. It showed that the overall incident rate decreased 0.52 incidents per year per 1,000 mile years of pipeline operation during the study period. The resulting R squared for this regression was 0.39. R squared values range from zero to one. They can be interpreted as the proportion of the variation in a given sample which can be explained by the resulting linear equation; they are a comparison of the estimated systematic model with the mean of the observed values. Very simply put, the closer the R squared value is to unity, the higher the relevance in the results. A similar regression was performed for external corrosion leaks only during the ten year study period. It indicated that the incident rate for external corrosion leaks was decreasing at the rate of 0.21 incidents per year per 1,000 mile years of pipeline operation during the study period. The resulting R squared was 0.24. 48

California State Fire MarshalJuLy 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-2A Incident Rates By Year Of Study Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile Years~YALCause of Incident 1981 1982 1198311984 1985 .~ 1988 1987 1:19881989 .~1990External Corrosion 4.78 7.21 4.19 3.36 3.14 3.73 5.673.952.893.55Internal Corrosion0.00 0.45 0.30 0.15 0.14 0.40 0.530.000.000.003d Party-Construction1.08 2.40 0.60 0.15 1.43 0.67 0.660.790.790.533rd Party-Farm Equipment3rd Party-Train Derailment1.08 0.15. 0.90 0.00 0.00 0.13 0.130.130.130.000.00: 0.00 0.00 0.00 0.00 0.00 0.000.130.130.003rd Party - External Corrosion0.00. 0.00. 0.00 0.00 0.14 0.00 O.130.00.0.000.663rdParty-Other0.15 0.30 0.60 0.00 0.14 0.40 0.00:0.13;0.13.0.13Human Operating Error0.31. 0.30 0.15 0.00 0.00 0.00. O.13~0.000.260.00Design Flaw0.00 0.00. 0.00 0.00 0.00 0.00 0.00~0.130.000.13Equipment Malfunction0.15 0.60 0.45 0.15 0.43 0.00: 0.400.920.260.39Maintenance0.00. 0.00 0.00. 0.00 0.29 0.00 0.00:0.00;0.390.00Weld Failure0.15 0.60. 0.60 0.29 0.43 0.13 0.130.260.000.13Other0.46 0.45: 0.15 0.29 0.29 0.67: 0.130.390.530.13Total8.18 12.47~ 7.94 4.39 6.42 6.13. 7.916.84~5.525.65Numberof Mile Years6.482 6,658: 6,675. 6,835 7,005 7,501 7.5877,600:7,6097,610Mean Year Pipe Constructed1952 1953: 1953~ 1954 1954 1956. 19571957~1957:1957Mean Operating Temperature F 97.0 97.4: 97.4 96.8 98.4 97.9 98.0 Mean Diameterinches 10.6 10.9; 10.9 10.9 11.1 12.3 12.3 Average Spill Size barrels 285.0 514.7 889.3 83.6 562.9 609.4 266.6;Average Damage $1,000US 1994 16.4 39.4 138.0 38.1 140.4. 255.7 31.8: Incident Rates By Year of Study Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile Years97.912.4136.290.398.0.12.4.377.5;968.698.012.4127.4210.315105- Extemal Corrosion1984 ~ 1985 1986Third Party All1987 1988`-~ Operat1989 1990ing ErrorWeld FailureEquipment MalfunctionOther49

~.Á ~ California State Fire Marshal July 1996STA~RREMAKSWJ. Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The decreasing trend in incident rates is especially noteworthy considering the fact that all leak data was gathered at the end of the study period. With the increasing trend towards total leak reporting and recording, one would assume that the more recent data collected from a pipeline operator may be more complete than data regarding leaks which occurred several years ago. This would tend to result in relatively lower incident rates for early study years and a corresponding increasing incident rate trend. However, as discussed earlier, the data indicated a rather significant decreasing incident rate trend. This indicates two things: first, it indicates that the data gathered is likely relatively complete during the earlier years of the study; secondly, it indicates that if any incomplete record keeping did occur during the early years of the study period, the actual decreasing incident rate trend was higher than indicated by the regressions. To reiterate, the data indicated a rather significant decreasing incident rate trend, which may actually have been somewhat understated. A third regression was performed for leaks caused by all causes except external corrosion during the ten year study period. It indicated that the incident rate for these leaks was decreasing at the rate of 0.19 incidents per year per 1,000 mile years of pipeline operation during the study period. The resulting R squared was 0.26. The average spill volumes varied widely during the ten year study period. An ordinary least squares line of best fit was determined to analyze any trend in this data. It indicated a 33.6 barrel per year reduction in average spill size, with an R squared of only 0.16. Finally, ordinary least squares lines of best fit were determined for the average cost of damage per incident during the ten year study period. Prior to running the regressions, all cost data was normalized to constant 1983 US dot lars~ Using all incidents during the study period yielded a $33,040 $US 1983, $49,145 $US 1994 per year increase in average spill cost, with an R squared of 0.27. After deleting the 1989 San Bernardino train derailment, the regression indicated a $23,366 $US 1983, $34,755 $US 1994 per year increase in average spill cost, with an Rsquared of 0.33. California Crude Oil Pipelines Under Study Table 4-2B presents leak incident data for California's crude oil pipelines under study, by year, during the three year study period. The data sample indicates a sharp increase in the frequency of incidents per year. However, nearly all of the 1995 leaks occurred on one line, which the operator plans to replace. This situation points out the severe limitations of the very small three year data sample; this sample precludes the meaningful analysis of any trends which might exist. We recommend that an analysis, similar to that conducted for the regulated California hazardous liquid pipelines, be conducted after several years of additional data has been collected. 50

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-2B Incident Rates By Year of Study California Crude Oil Pipelines Under Study Incidents Per 1,000 Mile YearsSTAT!RE~1~HALCause of Incident1993 1994 1995: -External Corrosion0.00 6.056.05. Internal Corrosion 2.02 0.00 3rd Party - Construction 0.00 2.023rd Party - Farm Equipment 0.00 0.00 Total 2.02 8.06 Number of Mile Years 494 4962.020.002.0210.08496Average Spill Size Barrels 1.0 295.5Average Damage $US 1994 : $5,000 $92,750 Incident Rates By Year of Study Incidents Per 1,000 Mile Years 7.6$2,840121086420:External Corrosion.Internal Corrosion3rd Party - Construction:3rd Party - Farm Equipment1993 1994 199551

cI.~. 0 California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 4.3 Decade of Construction Effects Regulated California Hazardous Liquid Pipelines The 1993 study regarding the state's regulated hazardous liquid pipelines concluded that pipe age had a definite effect on the leak incident rates. Table 4-3A shows the variation in leak incident rates by decade of pipe construction for these regulated pipelines. As indicated, pipe construction before 1940 1926 mean year of construction had a leak incident rate nearly twenty times that of pipe constructed in the 1980's. An ordinary least squares line of best fit was determined to evaluate the statistical relevance of the overall leak data by year of pipe construction. It indicated that the overall leak incident rate decreased 0.286 incidents per year per 1,000 mile years. The resulting R squared for this regression was 0.82. A second regression was performed which excluded all pipe installed prior to 1940. This regression indicated an overall leak incident rate reduction of 0.147 incidents per year per 1,000 mile years, with an R squared of 0.86. The study indicated that the vast majority of the difference in leak incident rates occurred because of variations in external corrosion rates. Some of the reasons for this variation may have included: The extent of external corrosion is generally considered a function of time. In general, the more time a given portion of pipe is allowed to corrode, the more likely it will be to develop a leak. Most believe that modern coatings are generally more effective than older coatings, especially those installed before the 1940's. The older pipe is likely to experience a higher external corrosion incident rate as a result. External corrosion rates are generally higher at elevated temperatures. Prior to the 1950's, it was common to install pipelines with little or no cathodic protection. For the most part, these older systems have either had new systems installed, or their older systems upgraded, to be consistent with present day practices. However, they often operated for several years with inadequate or no cathodic protection. The corrosion which occurred during these early years likely increased the resulting external corrosion leak incident rate. 52

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesSTATk!~.~_J L. Table 4-3A * Incident Rates By Decade of Construction Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile Years* `Cause of InddentI Pre-1940 1940-49 1950.59 1 1960~69 I: 1970-79:'~.:1930~9'E,rtemalCo~osion:' `14.12' 4.24 2.47 1.47 1.240.00*::jntemaICo,.~jon..0.38 0.27 0.10. 0.16 0.000.283rd Party - Construchon1 96 1 06 0 68 0 66 0 25028`1: 3rdP‡ryI.~FarmEquipmenti ~ 0.53 1.33 0.05 0.00 0.00 :3rd Party ~`Train Derailment.. ;I:~.~II 0.00 0.00 0.00 0.05 0.250.000.00 `~ :3~ Pa E~rnal'Co~ˆsjoh'"H 0.45 0.00 0.10: 0.33 0.000.003rd Party Other 0 30 0 13 0 05 0 05 0 000 000.30: 0.13. 0.O0~ 0.11 0.250.000.08. 0.00 0.00: 0.00 0.000.14*:Equipment.Malfu rtction: * . `1 0.38 0.53 0.10: 0.60 1.240.00Maintenance 0 00 0 00 0 24 0 00 0 000 00*:`"WËld.Failure'± ` :::: 0.38. 0.27 0.15. 0.44. 0.25~0.00: .~`Other:: `~" "`.` `..`1 0.83 0.13 0.24: 0.27 0.250.28 ::`T'otal, I.'~. ~. .1 :,~,,I 19.70. 8.08 4.17 4.15. 3.72:0.97NumberofMileYears 13247 7546 20612 18311 40307252AverageYearPtpe Constructed 1926 1944 1955 1965 19741985*:::Average:OpethtingTemperature:!F ~ 125.2 79.7 89.4: 91.4 99.8 AverageDiameterinches :, ~ 8.58 11.11 11.82 11.27 13.79104.119.55Il :`.`.:.:AveragespilrSize~barrels: . `~::.~I 162, 492 246~ 1,306: 53789..Ave~eDa~gË~$US 1994 : ~, H 46.517 177,902 252,479' 738.001 127.589244.407Incident Rates By Decade of ConstructionRegulated California Hazardous Liquid PipelinesIncidents Per 1,000 Mile Years25~.~*,2015 ~*~: ~ :~1050k- Pre-1940 1940-49 1950-59 1960-69 1970-79 1980-89- External Corrosion ~ Third Party All Operating Error~ Weld Failure Equipment Malfunction ~ Other-53

~4.California State Fire Marshal July 1996ST~ Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines An ordinary least squares line of best fit was determined for the external corrosion data only. Using all data, it indicated that the external corrosion rate declined by 0.217 incidents per year per 1,000 mile years, with an R squared of 0.79. A similar regression was performed excluding all pipe constructed prior to 1940. This regression indicated an external corrosion rate reduction of 0.097 incidents per year per 1,000 mile years, with an R squared of 0.95. However, it should be noted that both of these regressions resulted in a least squares line fit which would indicate a negative incident rate during the study period, which is impossible. However, the point should be made that there is a strong statistical relationship between pipe age and rate of external corrosion; the newer the pipe, the lower the external corrosion incident rate. A third ordinary least squares line of best fit was prepared for leaks caused by all causes except external corrosion. It indicated that the incident rate for these leaks decreased at the rate of 0.069 incidents per year per 1,000 mile years. The resulting Rsquared was 0.80. California Crude Oil Pipelines Under Study While the regulated California hazardous liquid pipeline data indicated a very strong correlation between pipe age and leak incident rates, we did not find the same correlation for the crude oil pipelines included in this study. Table 4-3B presents the leak incident rates by decade of pipeline construction. As shown, there is little correlation between pipe age and the incident rates for these pipelines. The oldest group of pipe, that constructed before 1940, had a leak incident rate of 2.21 incidents per 1,000 mile years. The group with the highest leak incident rate was constructed in the 1960's; this group had a leak incident rate of 16.95 incidents per 1,000 mile years. Similar to the analysis for the regulated California hazardous liquid pipelines, an ordinary least squares line of best fit was used to evaluate the statistical relevance of the overall leak data, by year of pipe construction, for the crude oil pipelines under study. It indicated that the overall leak incident rate was decreasing at the rate of 0.030 incidents per 1,000 mile years, for each year of decreasing pipe age. However, the resulting R squared for this regression was only 0.01, indicating little, if any, statistical relevance to this data. A similar regression was performed for external corrosion leaks only. This analysis indicated that the external corrosion leak incident rate was decreasing at the rate of 0.10 incidents per 1,000 mile years for each year of decreasing pipe age; the R squared for this regression was 0.14. As a result, the data for the crude oil pipelines under study does not indicate a statistical correlation between pipe age and the resulting leak incident rate. We suspect that this is largely due to the limited data sample available for this study. With a larger data sample, we would anticipate results similar to those for the regulated California hazardous liquid pipelines, for the same reasons discussed at the beginning of this section. 54

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-3B Incident Rate By Decade of Construction California Crude Oil Pipelines Under Study Incidents Per 1.000 Mile YearsSTA1EFIRE I.Cause of IncidentPre-1940 194O~i949~i950-i959~196O-1969~19704979 [..198O#989:~199O41 995ExtemalCorrosion2.21 13.09 0.00 11.30 0.00 0.00 0.00Internal Corrosion0.00 0.00 4.97 5.65 0.00 0.00 0.003rd Party - Construction0.00 0.00 0.00 0.00 0.00 5.38 0.003rd Party - Farm Equipment0.00 0.00 0.00 0.00 0.00 0.00 0.00Total2.21 13.09 4.97 16.95 0.00 5.38 0.00Number of Mile Years451.9 229.1 201.0 177.0 94.4 185.9 21.1Mean Year Pipe Constructed1930 1945 : 1954 1967 1974 1985 1992Aean Operating Temperature °F54.3 76.5 70.4 78.2 92.2 102.2 137.1Mean Diameter Inches7.8 6.6 6.1 10.6 6.5 7.4 9.6Average Spill Size Barrels4.0 3.3 25.0 1.7 0.0 589.0 0.0Average Damage SUS 1994$5,000 $6,067 $5,000 $3,333 $0 : $176,000 $0 Incident Rate By Decade of Construction Incidents Per 1,000 Mile Years20151050- External Corrosion 3rd Party - Construction Internal Corrosion~ 3rd Party - Farm EquipmentPre-1940 1940-1949 1950-1959 1960-1969 1970-1979 1980-1989 1990-199555

California State Fire Marshal July 1996STAThflREM~J4AL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 4.4 Operating Temperature Effects Regulated California Hazardous Liquid Pipelines The 1993 regulated California hazardous liquid pipeline study concluded that pipeline operating temperature had a definite effect on the leak incident rates. Table 4-4A shows the variation in leak incident rates by operating temperature for these regulated California hazardous liquid pipelines. With the exception of the relatively new pipelines operating at above 180°F most were built around 1979, higher operating temperatures were directly related to higher leak incident rates. However, the data also indicated that the pipelines operated between 130 and 159°F were also the oldest. As a result, a logistic regression was performed to determine whether or not pipe age was masking the pipe operating temperature effects. The logistic regression results indicated that while holding various factors constant, including pipe age, operating temperature was positively related to the probability of a leak occurring from external corrosion. Operating temperature was not statistically related, however, to the probability of leaks occurring from other causes. Ordinary least squares lines of best fit were also calculated to evaluate the statistical relevance of this regulated California hazardous liquid pipeline data. For all leaks, the line indicated an increase of 0.11 incidents per 1,000 mile years, per °F increase in operating temperature, with an R squared of 0.89. For external corrosion leaks only, the regression resulted in an increase of 0.10 incidents per 1,000 mile years, per °F increase in operating temperature, with an R squared of 0.91. For all leaks, excluding external corrosion leaks, the regression resulted in an increase of 0.0077 incidents per 1,000 mile years, per °F, with an R squared of only 0.28. These data reaffirm the logistical regression results that the probability of leaks occurring from external corrosion was affected by operating temperature, while leaks from other causes were not affected by operating temperature. The regulated California hazardous liquid pipeline data also indicated that spill sizes and monetary damage did not appear to be affected by operating temperature. California Crude Oil Pipelines Under Study The data for California's crude oil pipelines in this study did not indicate a similar operating temperature versus leak incident rate relationship. As shown in Table 4- 4B, there was no correlation between operating temperature and the leak incident rate associated with California's crude oil pipelines. sc

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesSTA1I'1R~MAJ~MAL Table 4-4A Incident Rates By Normal Operating Temperature Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile YearsCauseof Incident r 0-69°F :70~.99°F ~:100-i29°FL130..159°F,F~160°F,.+::ExternalCorrosion 0.48 1.33 7.11 11.3611.31Internal Corrosion - 0.00 0.21 0.32 0.570.083rd Party - Construction 1.91 0.94 0.95 0.570.603rd Party -Farm Equipment 0.00 0.30 0.47 0.000.083rd Party-Train Derailment 0.00 0.04 0.00 0.000.003rd Party - ExtemaI~Corrosion 0.00 0.06 0.16 0.00:0.153rdParty-OtherH: 0.00 0.24: 0.16: 0.000.15Human Operating Error 0.00. 0.11 0.00 0.000.23Design Flaw 0.00 0.04. 0.00 0.00:0.00Equipment Malfunction 0.00 0.24. 0.16 0.570.98Maintenance I 0.00 0.09 0.16 0.00.0.00Weld Failure 0.00 0.19: 0.32 o.oo:0.60Other0.00. 0.21. 1.11 1.140.45Total2.38, 4.01 10.90 14.2014.63Number of Mile Years2,097 46,641 6,332 1,760~13,260Mean Year Pipe Constructed1960 1959. 1953 19471951Mean Operating TemperatureF61.66 74.72 103.37 144.84~177.63Mean Diameter inches8.62 12.58. 11.88 9.9212.96Average Spill Size barrels12 480 72 7~601Average Damage SUS 1994 72,002 363,891 .53.866 15,566 Incident Rates By Normal Operating Temperature Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile Years14259020151050 0-69°F 70-99°F- External Corro~on100-129°F 130-159°F 160°F+Third Party AllOperating ErrorWeld FailureEquipment Malfunction ~ Other57

STATE FIRE MARE~4M. California State Fire Marshai July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-4B Incident Rates By Normal Operating Temperature California Crude Oil Pipelines Under Study Incidents Per 1000 Mile Years~:Causeoflncident~:. J0.69'F ~External Corrosion6 86 0 00 0 00 0 00 0 00lntema~CoiioÒ2.29 0.00 0.00 0.00 0.000.00 0.00 21.28 0.00 0.003rd Paity~ FarmEguipmeÒt.1.14 0.00 0.00 0.00 0.00Total1030 000 2128 000 000874.0 166.0 47.0 34.0 124.0Mean~YearPipËCˆnstiucted:~.~.1948 1961 1977 1987 196253.1 83.9 109.1 147.0 177.28.0 5.8 5.8 10.3 7.75.2 0.0 1174.0 0.0 0.0Average Damage $US 1994 $4 467 $0 $350 000 $0 $0- External Corrosion 3rd Party - Construction Internal Corrosion~ 3rd Party - Farm EquipmentIncident Rates By Normal Operating Temperature California Crude Oil Pipelines Under Study Incidents Per 1.000 Mile Years2520151050f,.0-69°F70-99°F 100-129°F 130-159°F 160+°F

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~UEMA1tS~M. An ordinary least squares line of best fit was used to evaluate the statistical relevance of the overall leak data, by operating temperature, for the crude oil pipelines under study. It indicated that the overall leak incident rate was increasing at the rate of 0.06 incidents per 1,000 mile years per 1°F increase in operating temperature. However, the resulting R squared for this regression was only 0.08, indicating little statistical relevance to this data. A similar linear regression was also performed on the external corrosion caused incidents only. This analysis resulted in a decreasing external corrosion incident rate of 0.04 incidents per 1,000 mile years, per 1°F increase in operating temperature. The R squared for this regression was 0.48, again indicating little statistical relevance to this data. It's also worth noting that all six of the external corrosion caused incidents occurred on pipelines operating in the ambient temperature category. This group was the largest, comprising 70% of the pipe sample. It was also the oldest pipe, with a 1948 mean year of pipe construction. See also Section 4.3 of this report for a discussion of pipe age effects. For the crude oil pipelines under study, these results do not indicate a statistical correlation between elevated pipe operating temperature and any increased risk of leak incidents. However, one must keep in mind the limited size of this data set. The small number of leaks 10 included in this limited three year study period, with only 496 miles of pipelines, is a very small sample. As noted earlier, this sample may not be large enough to show trends. 4.5 Pipe Diameter Effects Regulated California Hazardous Liquid Pipelines For the regulated California hazardous liquid pipelines, the leak incident rate for pipe 7" in diameter and less was over three times that for pipe larger than 20" in diameter 10.35 versus 3.17 incidents per 1,000 mile years. This is especially noteworthy since the mean operating temperature for the small diameter pipe was only 77.9°F, the lowest of any diameter range. However, the age of pipe in this category and in the 8-10 inch category was fairly old, which would tend to result in higher incident rates, as shown in earlier sections. This data is also presented in Table 4-5A. The category of pipe in the 11-15 inch diameter range also had a relatively high incident rate 8.62 incidents per 1,000 mile years. Although these lines were a good deal newer, they operated at a higher mean operating temperature. Surprisingly, the 16-20 inch pipe diameter range had a relatively low leak rate 3.49 incidents per 1,000 mile years, despite having the highest mean operating temperature range. 59

STAT MAP~HAI. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-5A Incident Rates By Pipe Diameter Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile YearsCauseoflncidentQ-7" .8-10" ~ 11_15' i 16-20" :1.:2O~.,-.External Corrosion6.75 4.56: 5.51 1.310.40Iinternal Corrosion0.33 0.27 0.13 0.070.00.3rd Party - Construction :1.96 0.83 0.97 0.360.793rd Party - Farm Equipment0.33 0.27 0.00 0.510.003rdParty-Train.Deraiiment :0.00 0.00 0.06 0.070.003rd Party- Extem‡iCorrosiÛn0.22 0.13 0.06 0.000.00.1.1.: . 3rd Party-Other .0.00 0.20 0.45 0.07~0.00.:: ...Human Operating Error...... . 0.11 0.10 0.26 0.000.00.. . Design Flaw....... : 0.00 0.03 0.00 0.000.40. . Equipme~*Maifunction: : :0.44 0.17 0.58: 0.361.19..... . Maintenance . . .0.00 0.03 0.06 0.150.00IWeld Failure :..:,0.00 0.30 0.261 0.3610.40 1 Other . : .0.22 0.57 0.26 0.220.00. . . . Total . . . . .10.35 7.46. 8.62 3.49:3.17. Numberof Mile Years ..9,183 30021 15,435 13,760.2,525MeanYear Pipe Constructed. :..1951 1948 1962 19641984Mean Operating Temperature "F.....77.90 94.11 104.81 108.44.91.17.Mean Diameter inches.....5.6 8.7 12.6 17.629.4Average Spilt Size barrels....55 190 489: 1,980.88:Average Damage $US 1994 . 26.981 93735 643.141 194.567 Incident Rates By Pipe Diameter Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile Years526. 7881210864200 - 7" 8 - 10"11 - 15" 16-20" 20"+- External CorrosionThirdParty AllOperatingErrorWeld FailureEquipmentMalfunctionOther60

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STArE~~R~ The largest pipe, over 20 inches in diameter, had the lowest leak incident rate, 3.17 incidents per 1,000 mile years. However, this pipe was the newest of any category, with a mean year of pipe construction of 1984. The mean operating temperature was moderate. Three ordinary least squares lines of best fit were prepared using this data. The first, performed using all data, indicated an overall reduction in the leak incident rate of 0.29 incidents per 1,000 mile years, per diameter inch increase, with an R squared of 0.76. The second, included only external corrosion leaks; it indicated a reduction of 0.26 incidents per 1,000 mile years, per diameter inch increase, with an R squared of 0.82. The third was performed using all leaks except external corrosion caused leaks; it resulted in a reduction of only 0.03 incidents per 1,000 mile years, per diameter inch increase, with an R squared of 0.31. In short, for the regulated California hazardous liquid pipelines, there was a correlation between pipe diameter and the incident rate for external corrosion leaks, but not for leaks caused by other factors. There are several possible explanations for this correlation: Larger diameter pipelines represent a larger capital investment for the pipeline operator. As a result, there may be a greater proportion of the operators' resources directed toward their construction, operation, and maintenance. * The larger diameter lines are often more important to the operators' overall operation and/or revenue generation. As a result, they may receive more attention. The larger lines are likely to create a greater perceived risk in the event of their rupture. This could also cause an operator to direct more resources to their protection. California Crude Oil Pipelines Under Study Slightly more than 90% of California's crude oil pipelines under study are 10" or less in nominal diameter; roughly 50% of the lines are 7" or less in diameter. Table 4-5B presents the incident rates and distribution by pipe diameter range. A statistical analysis was performed to examine any relationship between pipe diameter and the resulting leak incident rate for these pipelines. Somewhat surprisingly, a statistical relationship was not found for this limited sample. 61

$TATk~~ MAI~HA~. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-5B Incident Rates By Pipe Diameter California Crude Oil Pipelines Under Study Incidents Per 1000 Mile YearsCause;of:lncident .~0-7" 8-10" 11-15" .16-20"~ .~.20"+External Corrosion1.32 8.25 0.00 0.00 0.00- interrialCorrosion .1.32 1.65 0.00 0.00 : 0.003rd Part~Cˆnstruction0.00 0.00 13.33 0.00 0.003rd_~y-F~rrn~Eguipment .~0.00 1.65 0.00 0.00 0.00Total ~ ..2.65 11.55 13.33 0.00 0.00Number of Mil.:y~~ : .756 606 75 7 44Mean Year Pipe ConStructed...1955 1947 1968 1976 1970Mean OperatingTemperature'F :76.0 72.8 83.9 60.2 67.1Me·nOiamËtËlnches1 .: ~ .5.5 8.3 11.4 16.0 22.1Average SpillSizeBarrels14.0 2.7 1174.0 0.0 0.0Average D~m·ge $US 1994 . .$7,500 $3,600 $350000 $0 Distribution of Pipe By Pipe Diameter California Crude Oil Pipelines Under Study 3.0% 20"+ 0.5% 16 - 20" 5.0% 11 - 15"50.8% 0- 7"

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines151050 Incident Rates By Pipe DiameterCalifornia Crude Oil Pipelines Under Study Incidents Per 1000 Mile Years16- 20"SrATh FIRE L- External Corrosion 3rd Party - Construction Internal Corrosion~ 3rd Party - Farm Equipment0-7" 8-10" 11-15"*~--- _~r~1II20"+63

.~Q. California State Fire Marshal July 1996 StATh FIRE ~ Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Two ordinary least squares lines of best fit were prepared using this data. The first, performed using all data, indicated an overall reduction in the leak incident rate of 0.50 incidents per 1,000 mile years, per diameter inch increase; however, the resulting R squared was only 0.26, indicating little statistical relevance. The second, analysis included only external corrosion leaks; it indicated a reduction of 0.26 incidents per 1,000 mile years, per diameter inch increase, with an R squared of only 0.23. In short, for the California crudeoil pipelines under study, there was not a correlation between pipe diameter and the resulting leak incident rate. 4.6 Leak Detection Systems The California crude oil pipelines under study and the regulated California hazardous liquid pipeline data were sorted into pipelines having some type of supervisory control and data acquisition SCADA systems, and those without. These data are presented in Tables 4-6A and 4-6B, for the regulated California hazardous liquid and the crude oil pipelines under study respectively. In the 1993 study, 85% of the regulated California hazardous liquid pipelines had SCADA systems. For California's crude oil pipelines under study however, only about 9% of the pipelines had some sort of SCADA system installed. This difference is depicted graphically in Table 4-6C. For the crude oil pipelines under study, the leak incident rate for pipelines without these types of systems was roughly the same as the incident rate for systems with SCADA, 6.80 versus 6.13 incidents per 1,000 mile years. For the regulated California hazardous liquid pipelines, the pipelines with SCADA had a lower incident rate than those without, 6.29 versus 11.0 incidents per 1,000 mile years. However, this does not indicate that SCADA systems reduce leak incident rates. The average spill size and property damage was much larger for the crude oil pipelines under study with SCADA, than those without 1174 versus 5.2 barrels and $350,000 versus $4,467 respectively. However, there was only one leak on the 54 miles of pipeline with SCADA and nine leaks on the 441 miles of pipeline without. As a result, the data set is too small to draw any meaningful conclusions. Although the data set was too small to be meaningful, the results are somewhat surprising. SCADA systems generally provide a means of detecting leaks quickly, minimizing spill volumes; yet the leak on the pipeline system with SCADA resulted in the largest spill volume included in the study. This situation was also noted in the 1993 study regarding regulated California hazardous liquid pipelines. 64

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-6A Incidents By Leak Detection System Regulated California Hazardous Liquid Pipelines Incident Rate Comparison Incidents Per 1000 Mile YearsSTAT2F1~ MAg~iAL With SCADA ..,...:, ..~ ...Wflhout SCADA ICauseofincident Number .~: Rate~ ~ ~:Number Rate:External Corrosion . 214 3.49Internal Corrosion 13 0.21 87 1 0.091 11 1.011 3. 0.281 0 0.001 2 0.181 3 0.28~ 0: 0.001 0. 0.001 6~ 0.551 0 0.001 5. o.46J3rd Party-Construction53 0.86:3rd Party Farm Equipment15 0.24.3rd Party-Train Derailment2, 0.03.3rd Party-External Corrosion 5 0.08:3rdParty-Other 11 0.18Human Operating Error 8 0.13Design Flaw 2 O.03~Equipment Malfunction 21 - Q~34: Maintenance 5 0.08 Weld Failure 14 - 0.23.: Other 23 0.37. - 2 0.181 Total Numberof Mule Years Mean Year Pipe ConstructedMean Operating Temperature F Mean Diameter Inches 386 6.29.61351 1952, 114 3 &~$. ~ 124 ~ --- 120 ii.ool109041945 I107 0 95Average Spill Size Barrels 476 7 ~ -157 6 -~,Average Damage$US 1994 $228,972 ~$82,129 ~Incident Rate ComparisonRegulated California Hazardous LiquidPipelinesIncidents Per 1,000 Mile Years.12 .~, ~10 ~ ~ -~2 ~86- External CorrosionThird Party AllOperating ErrorWeld FailureEquipment MalfunctionOther . With SCADA Wthout SCADA65

5, STATh FIRE MAJtSMAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-6B Incidents By Leak Detection System California Crude Oil Pipelines Under Study Incident Rate Comparison Incidents Per 1,000 Mile Years * *~* ~. *~. H I :WitfrSCAD~ ~ : . WithoijtSCDACause of Incident J Number Rate Number Rate0 0.00 6 453*:lntÈm‡liCÛrrosion ~Ii0 0.00 2 1.51~3rd~Party-CÛnstrucb....,i;I~.:. 1 6.13 0 0.001. :;3~:pa~r~..:F Equipmentl:,: *~0 0.00 1 0.76H H H :. 1 6.13 9 6.80H: Number:Ûf:Mi Years .1..163 H 1324.:,:Mean,Year.Pipe.Constru~ed::HHH1965 195189.0 61.1 .::Mean:DiametÍrln~hËs I12.4 7.0 H ~ L H Average Damage:$US 19941:...1174.0 5.2$350,000 $4,467 Incident Rates By Leak Detection System California Crude Oil Pipelines Under Study Incidents Per 1,000 Mile Years86420Wth SCADA* - External Corrosion Internal Corrosion 3rd Party - Construction~ 3rd Party - Farm EquipmentWithout SCADA

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesSTATe!t~E!~t~j~AL Table 4-6C SCADA Systems Installation Distribution Regulated California Hazardous Liquid versus Crude Oil Pipelines Under StudyT-~TiiIi~~_~-100908070605040302010CCUI-- With SCADA Without SCADACalif. Reg. Haz. Liq. Crude Oil Under Study67

ji California State Fire Marshal July 1996STATBEIRE MARS~IA~ Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 4.7 Cathodic Protection System As indicated in Table 4-lA, 60% of the leaks on California's crude oil pipeline systems under study were caused by external corrosion. Because of this fact, the effectiveness of cathodic protection systems and cathodic protection system inspections were evaluated. Regulated California Hazardous Liquid Pipelines Nearly 100% of the regulated California hazardous liquid pipelines were protected by either impressed current or sacrificial anode cathodic protection systems. We did not find a statistically relevant difference in the effect on leak incident rates between the two types of systems. However, we found a significant difference between protected and the few unprotected pipelines. As depicted in Table 4-7A, unprotected pipelines had an external corrosion leak incident rate over five times higher than protected lines. Although a small sample, the unprotected lines were much newer than those covered by a cathodic protection system. Unprotected lines also operated at a higher mean operating temperature and were smaller in diameter. Cathodic protection systems appear to reduce the frequency of pipeline ruptures due to external corrosion. Data was also collected regarding the frequency of cathodic protection surveys. Table 4-7B shows the overall and external corrosion only incident rates by the average frequency of cathodic protection surveys. Ordinary least squares lines of best fit were prepared to determine whether or not the frequency of cathodic protection surveys had any statistical relevance to leak incident rates. Surprisingly, the ordinary least squares lines of best fit showed a slightly decreasing incident rate with less frequent surveys. However, there was little if any statistical relevance to this data; the R squared values for all incidents and external corrosion only incidents were only 0.13 and 0.01 respectively. This situation may result from operators performing more frequent surveys on pipelines with higher leak incident rates. A multinomial logistic regression analysis was performed to analyze this parameter. It indicated that the frequency of cathodic protection surveys was not statistically correlated with the external corrosion leak incident rate. 68

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-7A Cathodic Protection System Regulated California Hazardous Liquid Pipelines Incident Rate Comparison Incidents Per 1,000 Mile YearsSTATBE MARSiAL *.,....Cause of Incident . ~,: ~Prote~ed Unes . ...; UnProte~ed LinesNo.:ÛflncidÈnts 1 . Incident Rate N of Incidents Incident Rate * External Corrosion . 1.~,:295 4.23~: 9 23.12Internal Corrosion .~..14~ 0.20~ 0 0.00 3rd Party -Construction .~.*: 64~ 0.92! 1 2.57 3rd Party FOrm.EguipmeÒt.~ . 18: 0.26: 0 0.00 3rd Party-Train Derailment ~ 2~ 0.03.1 0! 0.00:3rd Party - Exter1alCOrrO~ioI~:...~~J 5 0.07! 1 2.57 t.,..~. 3rdParty-Other:..~: ~ 11 0.16 31 7.71 ~ HumanOpe~ng.~Or .!...! 8~ 0.11~ 0~ 0.00 DesignFlaw. SI ~::..........;.J 2: 0.03;: 0 0.00.J..,.....EguipmentMaJftjn~io~ ~ * 27: 0.39:: o: 0.00.....*. M·ihtenance. ~.*. *:5~ 0.07~ 0.00WOId Failure . . .191 0.27 01 0.00 -...*-. * *.flthor + ****** Number ‡f Mile -Years. _______ Mean Year Pipe Co structed ... ~. _______ -~Mean Operatir~g Temperature ~........._________ .....Mean Diameter Inches .............____________ *__Average Spill_Size_Barrels--__-r ____________ ...Average Damage $US1994 -69

STATh~1RE MARShAL. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Incident Rate Comparison Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 MUe Years * External Corrosion ~ Third Party All r~ Operating Error ~Weld Failure Equipment Malfunction ~ Other3020100Protected Un-ProtectedIn

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesSTAflFIREMA~HA1. Table 4-7B Average Cathodic Protection Survey Interval During Study Period Regulated California Hazardous Liquid Pipelines Incident Rate Comparison Incidents Per 1,000 Mile Years Uptol;OYears 1:.1to2~0Ye·~. :2:tto.5.0.YeaNCause of Incident Total No i Rate Number Rate Number Rate:5..1.tolO.OYearsNumber RateExternal CorrOSion146 3.43 100 6.68 48 4.104 3.33IntemalCorrosion : ~10 0.24 4 0.27 0 0.00.0 0.003ttP·rty-CÛnstruction~ 146 1.08 9 0.60 6 0.51 1 0.833rd Party Farm Equipment10 024 7 047 1 0 090 0 00~::1H~3rd Pafly.~Train~DeraiIment . 1 0.02 0 0.00 1 0.09:0 0.00:3i~Part~-1ExterflatCorrosion3 0.07 0 0.00 3 0.26 1 0.833~.P~0ther~1 H ~1 :~ 9 0.21 4 0.27 1 0.09.0. 0.00Human Operating Error 6 0 14 2 0 13 0 0 000 000:~DÈsignF1aw1. :1 1 0.02 1 0.07 0 0.000 0.00Equiprneht.Malfunctjon.:. 21 0.49 3 0.20 3 0.260 0.00,i:MaintËnaÒcei~1:~ ~1 :::~ 5 0.12 0 0.00. 0 0.000; 0.00WÈldFajkireHH S :14 0.33 4 0.27 1 0.090. 0.00H:.: ~13 0.31 10 0.67 1 0.090 0.00:285 6.70 144 9.62 65 5.55.6 4. .:.j 42,524~J~~ 14,961 ~ 11,713~~~ Mean Year Pipe Constructed 1954 1958 1962 8Mean Operating Temperature F 93 3 98 5 ~ 98 1 Mean Diameter Inches 111 ~4!~Z~ 161 .~ ~ ii ~1,202~1953738 88 ~Incident Rate ComparisonRegulated California Hazardous Liquid PipelinesIncidents Per 1,000 Mile Years12 ~108 L~ ..,;~ -~H4~6:~j0Upto 1.0 Years1.1to 2.0 Years2.1 to 5.0 Years5.1 to 10.0 Years- External Corrosion .~ Weld FailureThirdEquipParty Allment Malfunction Operating Error~ Other71

California State Fire Marshal July 1996STA~E1REM.~SHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gatherinu Lines California Crude Oil Pipelines Under Study 64% of the crude oil pipelines included in this study are protected by cathodic protection system. 19% are unprotected. The data for the remaining 17% was either missing or unknown. This data is shown in Table 4-7C. A graphic comparison is also presented which compares the distribution of cathodically protected pipelines for both California's regulated hazardous liquid pipelines and crude oil lines included in this study. The leak incident rate for the crude oil pipelines under study was roughly 30% lower for cathodically protected lines than it was for unprotected lines 7.36 versus 10.80 incidents per 1,000 mile years respectively. Although the data set was small, this trend is consistent with the data presented for the regulated California hazardous liquid pipeline system. Table 4-7D presents the incident rates for the crude oil pipelines under study, which have cathodic protection systems installed. It differentiates between the leak incident rates for those systems which are regularly inspected, and those that aren't. The overall incident rate for the crude oil pipelines under study with cathodic protection systems that are regularly inspected was 9.24 incidents per 1,000 mile years, 32% lower than the protected lines which did not have regular cathodic protection system inspections. The data for external corrosion leaks only yielded a greater difference; the inspected systems had an external corrosion caused incident rate of 4.62 incidents per 1,000 mile years, less than one-half the external corrosion rate for un-inspected systems. 4.8 Pipe Specification Effects Another characteristic which could influence the propensity of leak incidents is the type of steel used in construction. Tables 4-8A and 4-8B present the incident rates for varying pipe specifications for the regulated California hazardous liquid and crude oil pipelines under study, respectively. Although different pipe specifications had varying incident rates, it must be recognized that other factors also affected these rates. Regulated California Hazardous Liquid Pipelines 78% of the regulated California hazardous liquid pipe are constructed of ASTMJAPI X grade material. Normally, this pipe is manufactured from relatively high quality steel, with more strictly controlled chemistry. The mean year of construction and mean operating temperature for X-grade pipe used in regulated California hazardous liquid pipelines were 1960 and 97.6°F respectively. 22% of the pipe was constructed of ASTM A53 material. The incident rate for this material was nearly 2.7 times higher than that for X-grade material. However, this pipe was on average 10 years older, which would tend to increase the incident rate. 72

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-7C Cathodic Protection System California Crude Oil Pipelines Under Study Incident Rate Comparison Incidents Per 1,000 Mile YearsProtected Lines Un-Protected LinesUnknownIncident Number~ Rate Number :Ra~e I*Corrosion 1 4 4.20 2 7.2000.001lntemalCorrosjon 1 1.05 1 3.600 0.00 .1 1 1.05 0 0.000 0.00 j - H 1 1.05 0 0.000 0.00 J7 736 3 1080I 952 278 Constructed 1952 1958OperatingTernperature F 71 7 895 Inches 82 7 3 Size Barrels 173 7 1 7 fSUS 1994 $54 314 .. $3 333 Cathodic Protection System Distribution Regulated California Hazardous Liquid versus Crude Oil Pipelines Under Study0 00025819476987 30 0$01009080~.7060f50403020protected10 unprotected~ unknown0 -__Reg. Calif. Haz. Liq. Calif. Crude Oil Under Study73

STATh MAX~HAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines California Crude Oil Pipelines Under Study Cathodic Protection System Incident Rate Comparison Incidents Per 1,000 Mile Years ;j External Corrosion Internal Corrosion ~ 3rd Party - Construction ~ 3rd Party - Farm Equipment External Corrosion ~3rd Party-All ~ Operator Error ~ Weld Failure Equipment Malfunction ~ OtherProtected UnprotectedRegulated California Hazardous Liquid Pipelines Cathodic Protection System Incident Rate Comparison Incidents Per 1,000 Mile Yearsa403020100~ - ProtectedUnprotectedIA

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-7D Incidents By Cathodic Protection System Inspections California Crude Oil Pipelines Under Study. Incident Rate Comparison Incidents Per 1,000 Mile Years_______________________ Inspected Not-Inspected Cause of Incident j Number j Rate Number Rate External Corrosion 3 4.62 3 10.17 Internal Corrosion 0 0.00 2 6.78 3rd Party - Construction 1 1.54 0 0.00 3rd Party - Farm Equipment 1 1.54 0 0.00 Total * Number of Mile Years 649 Mean Year Pipe Constructed 1959 Mean Operating Temperature SF 79*7 Mean Diameter Inches 8.6 Average Spill Size Barrels 1 237.4 Average Damage $US 1994 $74,040rloe: uniy catnoaicaluy protectea pipelines have been * - Extema~ Corrosion ~ Internal Corrosion 3: 3rd Party - Construction ~ 3rd Party - Farm EquipmentInspected vs. Not-Inspected Incident Rate Comparison Incidents Per 1,000 Mile Years20151050Inspected Not-Inspected75

STATe ~RE MARSHAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-8A Incidents By Pipe Specification Regulated California Hazardous Liquid Pipelines Incident Rate Comparison Incidents Per 1,000 Mile Years......X-Grade A53 andGrade,B. ~.Other*Causeoflncident .. : Number ..~.Rate~~ NumberRate ~:Number .j RateExternalCorrosion 871.80 1037.64841.7260.12 50.3700.00Construction :~0.70 130.96210.43Equipment. . 100.21 50.3700.00.-TrainDer‡ilment : 20.04 00.0000.00malcorrosion: 1 20.04 30.2200.00 . :. 110.23 10.0700.00OperatingError !~:: 30.06 20.15 .00.00 . :~: 00.00 10.0700.00Malfunction ::~: 160.33 90.6700.00Maintenance:. ~ ~ 20.04 10.0700.00Failure.....:~ . 140.29 40.30 .00.00 . ~:: ... 130.27 - 20.1515.21To~I..: : 20041314911.051157.36Mile Years 4841213489192Constructed 1960 1 * 1950~rn4~ ~1950~L$~Temperature F 97 6 1 85 367 1inches 13 18 81~~M'8 9Srze barrels 757 SUS 1994. $419,728 . 63 .: S162,473Z~: 24$49,082 r ~ ;~t~ Incident Rate Comparison Incidents Per 1,000 Mile Years 80~/r ~ 60 - 40 ~ 0 X-Grade A53 and Grade B- Exter~ Weldnal Corro~onFailureThirdEquipParty Allment Malfunctionr.~ Operating Error~ Other7~c

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-8B Incidents By Pipe Specification California Crude Oil Pipelines Under Study Incident Rate Comparison Incidents Per 1,000 Mile YearssTATk~zi:IvLCauseoflncrdentX+Grade :~c~andG,.~deBIOther :~:UpJciIˆwnExternal Corrosion0.00 15.530.001.11Internal Corrosion0.00 6.210.000.003rd Party -Construction3.82 0.000.000.003rd Party - Farm Equipment3.82 0.000.000.00Total7.63 21.740.001.11Number of Mile Years262 3223900Mean Year Pipe Constructed1978 195219551944Mean Operating Temperature F108.7 67.997.766.2Mean Operating Pressure 282.0 212.9 Mean Diameter Inches 11.0 5.9Average Spill Size Barrels . 588.5 5.7Average Damage $US 1994 $176,000 $4,743 Incidents By Pipe Specification Incidents Per 1,000 Mile Years62.86.50.0$0::~46.37.04.0$5,0002520151050~P ,X-GradeA53 and Grade BOtherUnknown- Extemal CorrosionInternal Corrosion:. 3rdParty - Construction3rd Party - Farm Equipment77

California State Fire Marshal`1, July 1996 MA~11AL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines However, the mean operating temperature was about 12°F lower, which would tend to reduce it. An extremely small sample of pipe fell into the miscellaneous or other category less than 1%. However, the leak incident rate for this sample was very high, nearly 14 times that of X-grade pipe. Although the pipe had a mean age nearly 10 years older, it operated at a mean operating temperature roughly 30°F cooler. California Crude Oil Pipelines Under Study Although the regulated California hazardous liquid pipelines were largely constructed of ASTMIAPI X-Grade pipe, with a small percentage of miscellaneous or other pipe material, the crude oil pipelines included in this study were just the opposite. 60% of the crude oil pipelines under study were constructed of unknown pipe specification material. 18% of the pipe was X-Grade material. The remaining 22% was either ASTM A53 or API 5L grade B pipe. It's interesting to note that the leak incident rate for the unknown pipe was by far the lowest - 1.11 incidents per 1,000 mile years, versus 7.63 and 21.74 for the ASTMIAPI X-Grade and ASTM A53/API 5L Grade B pipe respectively. The miscellaneous or other pipe was by far the oldest, with 1944 as the mean year of construction. However, this pipe was operated at the lowest mean operating temperature. Despite the large variation in the incident rates for these different pipe groups, the reader should note that the data sample was too small to support any meaningful conclusions. Further, although external corrosion caused the largest portion of the discrepancies, this disparity is likely caused by other factors; we do not believe that external corrosion is significantly affected by pipe specification. 4.9 Pipe Type Effects Regulated California Hazardous Liquid Pipelines Table 4-9A presents the regulated California hazardous liquid pipeline data by the type of pipe installed. The data sample was broken down into five categories: submerged arc weldedSAW, seamless SMLS, electric resistance welded ERW, lap welded LW and miscellaneous/other. The pipe included in this database was distributed as follows: 78

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-9A Incident Rates By Pipe Type Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile YearsIL~ t. * Cause of Incident SMAW IHSMLS1 ERW:L1~~LW~ OtherExtemalCorrosiori8.35 3.66 1.4731.590.00Internal Corrosion2.09 0.22 0.021.830.003rd Party - Con~truction0.00 0.86 0.456.410.003rd Party - Farm Equipment0.00 0.22 0.021.830.003rd Party - Train Derailment 0.00 0.00 0.020.000.003rd Party - External Corrosion 0.00 0.00 0.090.000.003rd Party - Other 0.00 : 0.00 0.120.460.00Human Operating Error . 0.00 0.11 0.051.370.00Design Flaw . 0.00 0.00 : 0.000.460.00Equipment Malfunction0.00 0.54 0.171.370.00Maintenance0.00 0.11 0.000.460.00Weld Failure0.00 0.00 0.121.830.00Other 0.00 0.43 0.142.290.00Total10.44 6.14 2.6849.900.00Number of Mile Years479 9,280 42,1122,1841,106 Mean Year Pipe Constructed 1978 1951 1963MeanOperatingTemperature~F 120.28 83.59 :~ 98.02 Average Spilt Size barrels 5 83 285 Average Damage SUS 1994 $28,008 $290,684 $602,431 Incident Rates By Pipe Type Incidents Per 1,000 Mile Years 1933 86.87 87$102,121195285.58 0 $060.~40ci~.~200SMAW SMLSERWLW OtherExternal CorrosionThirdParty AllOperating ErrorWeld FailureEquipment Malfunction ~ Other79

STATE FIRE MARSHAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesSeamless16.8%Lap Welded4.0%Submerged Arc Welded0.9%Miscellaneous/Other2.0%Pipe TypeElectric Resistance WeldedPercentage of Sample76.3% The data indicated that lap weld pipe had a very high leak incident rate; nearly 50 incidents per 1,000 mile years.. However, it was also the oldest pipe, with a mean year of construction of 1933. The weld failure caused incident rate for lap welded pipe was also the highest in the group 1.83 incidents per 1,000 mile years. Electric resistancewelded ERW pipe had a comparatively low incidence of leaks, 2.7 incidents per 1,000 mile years. These leaks occurred on somewhat newer pipeline systems, with a mean year of construction of 1963. They also operated at a mean temperature near the mean for the entire pipe sample. Seamless pipe experienced an incident rate of 6.1 incidents per 1,000 mile years. However, this pipe sample had a mean year of construction of 1951. The mean operating temperature was comparatively cool, 83.6°F. Submerged arc welded pipe had a high incidence of leaks, 10.4 incidents per 1,000 mile years. This small pipe sample had a mean year of construction of 1978. The mean operating temperature was the highest of the group, 120.3 °F. California Crude Oil Pipelines Under Study Table 4-9B presents the data for the crude oil pipelines under study. The pipe within this database was distributed as follows:Pipe TypePercentage of SampleElectric Resistance Welded22.2% approximately 110 milesSeamless11.9% approximately 59 milesLap Welded14.4% approximately 71 milesSubmerged Arc Welded2.9% approximately 14 milesUnknown46.9% approximately 223 milesMiscellaneous/OtherMiscellaneous/Othergo

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The data presented in Table 4-9B illustrates the limitations of this small data sample. Specifically, the miscellaneous/other pipe type, which includes drilling pipe, had the highest leak incident rate, 41.67 incidents per 1,000 mile years. However, this resulted from only one incident, caused by third party damage. Because of the very small inventory of pipe within this category, a very high incident rate resulted. As stated before, this data set is simply too small to provide meaningful analysis in many instances. The seamless pipe also had a relatively high leak incident rate 33.9 incidents per 1,000 miles years. This rate was nearly four times higher than that for the next highest pipe type ERW, with 9.06 incidents per 1,000 mile years. The biggest factor in this difference was external corrosion, which caused 28.2 incidents per 1,000 mile years for the seamless pipe, and 6.04 incidents per 1,000 mile years for ERW. Although this difference is large, external corrosion is not generally considered a function of pipe type. External corrosion is generally affected by pipe age, operating temperature, and other parameters. As a result, we do not believe that there is a correlation between pipe type and the leaks caused by external corrosion. This difference is likely caused by other factors and the small data sample available. The purpose of this evaluation was twofold: first, to determine the distribution of the crude oil pipe installed and second, to identify any explainable differencÁs in the leak incident rate caused by pipe type. While we were able to accomplish the first objective, we were unable to identify any link between pipe type and the resulting leak incident rate. 81

STAI~FIREMAI~SHAL California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-9B Incident Rates By Pipe TypeCalifornia Crude Oil Pipelines Under Study Incidents Per 1,000 Mile Years* P Cause‡fThcident :SMA'~~ ~ML II:. . ER~ Lw:: ~~:OthË~:I:.~PUhkiun0.00 28.23 6.04 4.67 0.00 0.00nternalCorr‡sion P;0.00 0.00 3.02 0.00 0.00 1.430.00 0.00 0.00 0.00 41.67 0.000.00 5.64 0.00 0.00 0.00 0.000.00 33.87 9.06 4.67 41.67 1.43Number of Mile Years 43 177 331 214 24 698 1969 1942 1972 1929 1985 1951::M~OrnrÈf~ 65.0 48.5 105.2 49.0 83.7 73.9Mean D~atneter lnches~ 22 0 6 4 7 0 8 3 6 3 6 9Average Spill Size Barrels 0 0 3 3 1 7 4 0 1174 0 25 0Average Damage $US 1994 $0 $5 050 $3 333 $5 000 $350 000 $5 000*Other category includes drilling pipe. .i~9T7~~ I ~* Ext 3rdernal CorrosionParty - Construction Internal Corrosion~ 3rd Party - Farm EquipmentIncident Rate Comparison Incidents Per 1,000 Mile Years50403020100jSMAW SMLS ERW LWP~Other* Unknown

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines SflR~S~1. 4.10 Operating Pressure Effects Regulated California Hazardous Liquid Pipelines The 1993 study concluded that the relationship between normal operating pressure and the probability of pipe rupture was not statistically significant. Table 4-IOA shows that there was considerable variance in the incident rate by pressure range. These differences, however, disappeared once variables such as age of pipe and operating temperature were controlled in the logistic regressions. A simple ordinary least squares line of best fit was also determined using the overall leak data for each pressure range. The data indicated a declining leak incident rate as operating pressure increased, with an R squared of 0.32. However, as indicated above, the logistical regressions, which take other factors into account, did not indicate a correlation between operating pressure and leak incident rates. An ordinary least squares line of best fit was also prepared for spill size as a function of operating pressure. The slope of the ordinary least squares line of best fit indicated a roughly 90 barrel increase in mean spill size per 100 psi increase in operating pressure. This regression resulted in an R squared of 0.62. It should also be noted that mean pipe diameter was also slightly higher for pipelines operating within the higher operating pressure ranges; this would also skew the results in this direction. A similar line of best fit was prepared for average damage as a function of operating pressure. The slope of the ordinary least squares line of best fit indicated a roughly $37,000 $US 1983, $55,035 $US 1994 increase in average damage per 100 psi increase in operating pressure. This regression resulted in an R squared of 0.58. However, as noted for spill volumes, pipe diameter variances would also generally affect spill damage. 83

STAT~FIP.E MAXS.HAI. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-IOA Incident Rates By Normal Operating Pressure Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile Years :0~ . 101_~,:.201_ :.i:.3o1 401- 501- I .601- :801~..1Q0i.;Incident 100 200 300 400 500 600 800 1000 psi .:.psi :1: :~si psi psi :psi .1: .~si psi .~ +ps~Corrosion 16.67 4.11 1.63 4.12 5.16 13.05 5.83' 1.261.58:. H 0.45 0.69 1.23 0.34 0.23 0.20 0.00 0.000.00Construction .. 1.80 2.29: 1.02 0.17 0.70 1.19 1.09 0.600.753rdParty-FarmEg~meflt~ H H 0.00 0.00 0.61 0.00 0.47 0.20 0.40 0.060.48Pa~~Trai~*Deraiirnent H: 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00:0.14 . 0.00 0.46 0.41 0.00 0.00 0.20 0.00 0.060.00Other . H 1.: 0.00' 0.69 0.41 0.00 0.00 0.00 0.10: 0.36 E~or I:..H 0.45' 0.00 0.20 0.00: 0.47 0.00 0.30; 0.000.140.07Flaw .. . I. 1:: 0.00 0.00 0.20 0.00 0.00 0.00 0.00 0.060.00Egui~mentMaffun~tion: 1.80 1.37. 0.00 0.17 0.00 0.00 0.69 0.300.21 : : 0.00 0.00 0.00 0.00 0.00 0.00 0.20 0.180.00Failure ~..: . 1.35. 0.00. 0.20 0.00 0.23 0.00 0.30 0.360.27Other: 0.90 0.46 0.20: 0.00. 0.70 1.19 0.20: 0.420.142343 1006 613 481 797 1601 910 365377,. :1 2,219 4,374 4,895 5,818 4,264 5,058 10,112:16.73214,597Pipe;Constructed . .: 1933 1954 1949 1940 1946 1934 1945 19581949AverageOpeating.Temperature°F:* 130.8 92.7 82.8 86.7 121.6 125.2: 159.7 116.2104.4:AverageD~rnet~r:lnches:.::.::::::: 9~9: 11.0 8.6 12.7 8.7 9.3 11.1 16.411.7S~eBa~1s *. .: :~ 17 56 5 130 149 127 456: 1,2926761994:.. 88 106 57 74 39 19 104 248872Incident Rates By Normal Operating PressureIncidents Per 1,000 Mile Years101-200 201-300 301-400 401-500 501-600 601-800 801-1000 1001+Normal Operating Pressure Range psiCorrosion ~ Third Party All ~ Operating ErrorFailure Equipment Malfunction ~ Other84

California State Fire MarshalJuly 1996 ~Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines sr* I~~M~S~MAL California Crude Oil Pipelines Under Study Not surprisingly, most of the crude oil pipelines under study were operated at relatively low pressures. In fact, 65% of these lines were operated at 100 psig or less. The operating pressure distribution for both the regulated California hazardous liquid and crude oil pipelines under study are presented below for comparison. Operating Pressure * Range psig . Crude Oil Pipelines. . Under Study . .Regulated California HazardÛ‡s Liquid . Pipelines- 100 psig64.7%3.3%101 - 200 psig8.2%6.4%201 -300 psig10.1%7.2%301 -400 psig9.7%8.5%401 - 500 psig2.2%6.3%501 - 600 psig3.3%7.4%601 -800 psig . 1.8%14.9%800+ psig0.0%46.0% As indicated in Table 4-lOB, there does appear to be a relationship between operating pressure and the resulting leak incident rate. Although we believe that leak incidents caused by third party damage are not related to operating pressure, it is reasonable to assume that operating pressure and leak incidents caused by internal and external corrosion could be related. Specifically, we found that the combined internal and external corrosion leak incident rates for crude oil pipelines under study were 26.00 and 18.21 incidents per 1,000 mile years for those operated between 201 - 300 psig and 301 - 400 psig respectively. The combined external and internal corrosion leak incident rate for pipelines operated at 100 psig or less was only 4.08 incidents per 1,000 incidents per 1,000 mile years. However, the pipe operated at higher pressures also operated at a higher mean operating temperature. But the pipe was generally newer, with a more recent mean year of pipe construction. Additionally, the lower operating pressure group of pipelines had the highest average spill size and average property damage. Although the data set was too small to draw any conclusions at this time, we believe that this parameter should receive additional consideration after several years of additional leak data has been gathered. 85

STATE FIRE MAREMAL California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oi~I and Crude Oil Gathering Lines Table 4-lOBIncident Rates By Normal Operating PressureCalifornia Crude Oil Pipelines Under Study Incidents Per 1,000 Mile Years Cause oflncid~nt:. ~0 101-200. 201~300: 0i~400.:.40l-500 :5O1~.60O1 601~~500 psig psig psig psig psig psig psigExtemalCorrosion~ 2.72 0.00 17.33 18.21 0.00 0.00 0.00~1ntemal Corrosion1.36 0.00 8.67 0.00 0.00 0.00 0.00~:3~dParty.Coi1structicn::::1.36 0.00 0.00 0.00 0.00 0.00 0.001.36 0.00 0.00 0.00 0.00 0.00 0.00Total I 6 79 0 00 2600 1821 0 00 0 00 0 00736 93 115 110 25 37 22f:;Meanyeai~:pjconsti.ucted 1945 1970 1958 1971 1979 1970 1971~MË‡nOpÈ9Te‡~pËmjureF 54.5 79.7 102.2 93.1 136.2 60.0 60.07.3 12.8 6.5 6.9 7.6 5.6 5.0242.0 0.0 1.7 3.5 0.0 0.0 0.0AveragiDam·ge~$tJS1994 I$74,000 $0 $3,333 $4,100 SO SO $0 30 20 10 0 Incident Rates By Normal Operating Pressure Incidents Per 1,000 Mile Years I _________ _________ _________101-200 201-300 301-400 401-500 501-600 601.800- External Corrosion Internal Corrosion:: 3rd Party - Construction ~ 3rd Party - Farm Equipmentor

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines s~ FrnEM.~SRAL 4.11 External Pipe Coatings This subsection examines the incident rates for various external pipe coatings. To accomplish this, the data sample was sorted into several categories, which represented nearly all of the coatings installed on the pipelines included in this study. These coating types, their common and trade names, and the percentage of each in operation during the study period are presented below. It should be noted that the coating type was reported as unknown on roughly 30% of the crude oil pipeline length included in this study. The figures below show the coating type distribution of the pipelines where the coating type was reported. .~~:.CrudeOi1 * Type Pipelines Under Study.______________ Regulated :~ * :, Cal~ornia Hazardous . . . ..Liquid :Pipelmes*~Common~~de: Names ~ :.Extruded with 15.6% Mastic6.5% X-Tru-Coat Plexco6OXT X-Tru-CoatBonded 2.7% .1.8% FEE MobiloxScotchcoat 206 or 202Thin Film Epoxywith 3.2%Butyl7.6%PritecAsphalt ~ 3%24 9% . SomasticAsphalt MasticSystems 0.0%41.6% ~Coal Tar EpoxyCarboline Epoxy .Applied 51%6.0% Polyken Tape YGI11 PlicoflexRaychem Hotciad SynergyTar 6.3%Coal Tar or AsphaltEnamel WrappedPipe 25.0%6.8%N/A25.8%0.0%N/A As indicated, there was a far greater percentage of bare pipe in the crude oil pipeline inventory under study than the regulated California hazardous liquid pipeline inventory 25% versus 6.8%. 87

California State Fire Marshal ; July 1996STAThFIRE ~ Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Regulated California Hazardous Liquid Pipelines Table 4-1 1A presents the leak incident rates by coating type for regulated California hazardous liquid pipelines. Although pipe age and operating temperatures had the greatest effect, there did appear to be differences in performance between the coating systems. The average external corrosion incident rate for the regulated pipelines was 4.18 incidents per 1,000 mile years. Generally, the more modern coatings had external corrosion incident rates lower than average, some significantly lower. The older asphalt mastic systems had slightly higher external corrosion incident rates. The coal tar and asphalt enamel wrapped pipe had an external corrosion incident rate nearly as high as the bare pipe. Bare uncoated lines, which comprised roughly 7% of the total, suffered the highest external corrosion and overall incident rates. In fact, these values were almost three times the average values for all pipelines included in the study. However, these lines had the oldest mean year of pipe construction and a mean operating temperature higher than average. The coal tar and asphalt enamel wrapped pipelines, about 5% of the total, had an external corrosion rate nearly as high as the bare pipelines. These lines were operated at an average of 8°F above the mean operating temperature. They were also on average 5 years newer than the mean. Extruded asphalt mastic coated pipe, roughly one-quarter of the total, had the third highest external corrosion and overall incident rates. This pipe had the second oldest mean year of pipe construction and the lowest mean operating temperature. The 2% of the total pipe coated with fusion bonded epoxy had the forth highest external corrosion and overall incident rates. The external corrosion incident rate for this coating was slightly below the overall average. This pipe was the newest sample included in the study, with a 1984 mean year of pipe construction. However, the operating temperature was the highest of the group, 115.6°F. Extruded polyethylene with asphalt mastic, liquid systems and mill applied tape had external corrosion incident rates roughly one-half to one-third the average. The overall incident rates for these coatings were also considerably lower than the average. The mean pipe age and mean operating temperatures varied considerably among these groups. However, the pipe was generally much newer than average, with higher than average operating temperatures. The lowest incident rates were observed on pipe with extruded polyethylene with side extruded butyl, which comprised 8% of the total. The observed external corrosion and overall incident rates for these pipelines were both less than one-tenth the average values. This pipe sample was relatively new, with a 1973 mean year of pipe construction. The mean operating temperature was moderately high, 105.8°F. 88

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines~ATe~REMA~ Table 4-hA Incident Rates By Coating Type Regulated California Hazardous Liquid Pipelines Incidents Per 1,000 Mile YearsExtrudedExtrudedCoalFEFusionFE withExtruded LiquidMillBare Tar orCause of IncidentwithBondedSideAsphalt SystemsAppliedPipe AsphaltAsphaltEpoxyExtrudedMasticTapeEnamelMastic FBE ButylAMWrappedExtemalCorrosion2.49. 3.71 0.36 5.56 1.27 158 11.77 11.59Internal Corrosion0.21 0.00 0.00 0.27 0.20 0.00 0.20' 0.293rd Party - Construction 1.04 0.00 0.18 1.31 0.49 0.45 1.60~ 1.453rd Party - Farm Equipment 0.42 2.22 0.00 0.22 0.00 0.45 0.00: 0.873rd Party-TrainDerailrnent 0.21 0.00 0.00 0.00' 0.03 0.00 0.00 0.003rdParty-ExtÈmalCorrosjon 0.00~ 0.00 0.00 0.16 0.13: 0.00~ 0.00 0.003rdParty-Other 0.21 0.00 0.00 0.16 0.16 0.23 0.80 0.00Human Operating Error ` 0.21' 0.00 0.00 0.11 0.07' 0.00 0.40~ 0.29Design Flaw ` 0.00~ 0.00' 0.00 ` 0.05 0.00. 0.23 0.00 0.00Equipment Malfunction 0.21 0.74. 0.00: 0.33. 0.33: 0.00 0.40 0.29Maintenance0.00 0.00 0.00. 0.11' 0.03. 0.00' 0.00' 0.00Weld Failure0.00. 0.00: 0.00 0.05: 0.20 0.68 0.40 0.29Other0.42' 0.74' 0.00 0.16 0.20 0.45 1.80: 0.58Total5.40 7.41 0.53 8.51 3.09' 4.06 17.35: 15.65 Number of Mile Years 4,814: 1,349 5,625 18,342' 30,700 4,435 5,013; 3,450 Mean Year Pipe Constructed , 1974 1984: 1973 1956 1959 1984 1948: 1962Mean Operating Temperature CF 107.4: 115.6 105.8 80.5 98.1 104.6' 103.8 105.8 Incident Rates By Coating Type Incidents Per 1000 Mile Years20151050PEw/AMPE w/Butyl A M Liquid Tape Bare Coal Tar- External Corro~ Weld Failuresion ~ Third Party All Operating Error Equipment Malfunction ~ OtherPE = Extruded PolyethyleneFBE = Fusion Bonded EpoxyAM = Extruded Asphalt MasticLiquid = Epoxy Liquid Applied SystemsTape = Mill Applied TapeButyl = Side Extruded Butyl Rudder89

`V California State Fire Marshal July1996srAazFIREM*.~MAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Difficulties were encountered performing multiple logic regressions using the coating type as an independent leak indicator. This occurred because the leak data and pipe data were gathered separately. Subsequently, the data were compiled using two separate databases. The coating type data was gathered for each segment of each pipeline within the State, resulting in tens of thousands of individual pipe segments. However, the leak data contained only the pipeline identification on which the leak occurred, as well as other pertinent data. The leak data did not specifically identify which segment of pipe suffered the leak. As a result, some manipulation of the data was necessary to perform the multiple logic analysis. The resulting analysis did indicate a correlation between coating type and leak incident rates. California Crude Oil Pipelines Under Study As noted earlier, 30% of California's crude oil pipelines under study were reported as unknown. 18% of the pipelines were bare. Another 18% was reported to be coated with other types of coatings. The next two largest groupings were extruded asphalt mastic /Somastic coated lines 11% and extruded polyethylene with asphalt mastic coated pipelines 11%. The remaining 12% of the lines were coated by a variety of coating systems. The highest leak incident rate was encountered on the coal tar or asphalt enamel wrapped pipelines. This result is consistent with the regulated California hazardous liquid pipeline data. The crude oil lines in this study which were coated with coal tar or asphalt enamel had a leak incident rate of 45.8 incidents per 1,000 mile years. However, this data sample was very small. The incident rate resulted from only three leaks on 22 miles of pipeline. Two of the three leaks were caused by external corrosion, resulting in an external corrosion caused incident rate of 39.5 incidents per 1,000 mile years. Two of the other external corrosion caused leaks occurred on pipe coated with somastic and other/unknown coatings. Only one external corrosion caused leak occurred on bare pipe. The external corrosion caused incident rates for the somastic, bare, and other/unknown coated lines were 5.85, 3.82, and 7.37 incidents per 1,000 mile years respectively. These rates are similar to the overall external corrosion caused incident rate for the entire California crude oil pipeline system included in the study - 4.02 incidents per 1,000 mile years. 90

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-IIB Incident Rates By External Coating Type California Crude Oil Pipelines Under Study Incidents Per 1,000 Mile Years `L.oaiiaror: Extruded PE I Fusion I AsphaltCause of Incident WI Asphalt Bonded Enamel Somastic Pritec Mastic AM Epoxy FBE i Wrapped CTExternal Corrosion 0.00 0.00 30.53 5.85 0.00Internal Corrosion 0.00 0.00 15.27 0.00 0.003rd Party - Construction . 0.00 0.00 0.00 0.00 0.003rd Party - Farm Equipment 0.00 0.00 0.00 0.00 0.00Total 0.00 0.00 45.80 5.85 0.00Numberof Mile Years164 28 66 171 34Mean Year Pipe Constructed1979 1978 1952 1955 1986Mean Operating TemDerature 1°F85.1 181.1 65.9 81.6 96.6 :~~.Causeoflncident : :. ~~i2i ~ Bare Pipe . Other UnknownSE taiCon~s~on, s: H 0.00 0.00 3.82 7.37 0.00Internal Corrosion 0 00 0 00 3 82 0 00 0 003rd Party Construct2on0 00 0 00 0 00 3 69 0 00531.65 0.00 0.00 0.00 0.00Total s .~.231.65 0.00 7.65 . 11.06 0.00NumberÛfMileYe·rs .:.. :1 32 21 262 271 440Mean Year Pipe Constructed 156 1953 1940 1959 1938 * .1 60.0 64.4 45.0 70.9 86.1Incident Rates By External Coating TypeIncidents Per 1,000 Mile Years50 -.~ --..---...14Q~ 230-20J .f]~t I.~ 10- A..................... ~ PE CT Pritec FAT Other FB Somastic MAT Bare Unknown - External Corrosion - Internal Corrosion 3rd Party - Construction 3rd Party - Farm EquipmentPE= Extruded Polyethylene with Asphalt Mastic MAT=MilI Applied TapeFB=Fusion Bonded Epoxy FAT=Field Applied TapeCT=CoaI Tar or Asphalt Enamel Wrapped91

California State Fire Marshal July 1996STA1~kE~.A~K4L Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 4.12 Internal Inspections During the last several years, there have been significant advances in the technologies available to internally inspect pipelines using instrumented devices commonly called "smart pigs ". These tools use several technologies to identify wall thinning, buckling, erosion, corrosion and other anomalies. These technologies, available from various vendors, differ greatly in their ability to identify and quantify various forms of pipe damage and/or deterioration. Some are precise and sophisticated, while others are much more general. Unfortunately, most of these inspection tools are rather long. As a result, they require smooth, long radius bends to facilitate their passage. Most will not traverse short radius elbows for example. In this Section, we will attempt to: * quantify the total length of pipelines which could be inspected using smart pigs, and * identify any differences in the leak incident rates for internally inspected pipelines. Regulated California Hazardous Liquid Pipelines Out of the roughly 7,800 miles of regulated California hazardous liquid pipelines, nearly 58% 4,495 miles are capable of being inspected using these techniques with little or no modification. 70% 3,128 miles of the pipelines which are capable of being inspected by smart pigs, have already been inspected in this manner. Table 4-12A presents a comparison of the incident rates for pipelines meeting three criteria: * pipelines which have been internally inspected, pipelines which could be inspected with little or no modification, but had not been inspected by the end of the study period, and those pipelines which are not capable of being inspected utilizing a smart pig without significant modification.

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-12A Incidents By Internal Inspections Regulated California Hazardous Liquid Pipelines Incident Rate Comparison Incidents Per 1.000 Mile YearsSTAT~E I - External Corrosion ~ Third Party All Operating Error ~Weld Failure Equipment Malfunction ~ Other .. .. Internally Inspectable .~ ` NOtIncident Inspected Not- Inspected Piggable s. *.... Number .1~ Rate :Numb~r ~. ~Rate ;Number, RateCo~sio~ ~ 65 2.0639 3.47198 6.70Co~Ùsion . ~, ~, 5 2 0.060 0.0012 0.413rdPary.ConstructiÛn':'::. 16 0.516 0.5342. 1 42Equ ipmÈnt t 8~ 0.250 0.00.10 034Derailment' 5 1 0.03 1 0.090 0 00PartyS-ExtemalCorrÛsion 2 0.060 0.005. 0.173rd.Party-Other~ ::,: `5 8 0.25 1 0.09:5 017.5;5H¸manOperatingError,~:~ 2 0.062 0.18:4 0 14`s s 1' 0.030' 0.00 1 003 : 12: 0.384 0.36:11' 0.37.....*: ., 3 0.10.0 0.00;2 007Weld:FajIure' : 5 11 0.350 ODDs8 0.27*: :5,~ :: . 8' 0.258 0.719 0.30:: ~ ~ 139 4.4161 542307 10.39Mile Years 31 500~~~11 253 ~29 550~ `~Total Mile Years 436% Miles 3 ~15 6%1 367 -~ `~409% ~3 305bTotal Length 40 1% ~17 5% `~1424% - Constructed ~Temperature SF 121 ~ inches . ~1941 ~. .~14813.019441 97 8.7~Incident Rate Comparison Incidents Per 1,000 Mile Years12:~q1086 ~42Inspected Not Piggable Inspectable - Not-Inspected93

California State Fire MarshalJuly 1996STATflFIREMAJSHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The data indicates that pipe which had been internally inspected had the lowest leak incident rate. However, this pipe was also the newest of any category, with a 1963 mean year of pipe construction, 6 years newer than average. This pipe was also operated at a mean operating temperature of 121°F, 23°F higher than average and had the highest mean pipe diameter, 15.3". We also compared the two categories of pipe which had not been internally inspected. Although the pipe which was not capable of being inspected using a smart pig was newer and operated at a lower mean operating temperature, it had an overall incident rate almost double the rate for piggable pipe which had not been inspected. However, the mean diameter for non-piggable lines was much smaller, 8.7" versus 13.0". California Crude Oil Pipelines Under Study Only 2% roughly 8 miles of the California crude oil pipelines under study have ever been internally inspected using a smart pig. Another 4% approximately 20 miles could be internally inspected, but has not been inspected in this manner. The remaining 96% 468 miles could not be internally inspected because of physical limitations e.g. short radius elbows. Table 4-12B presents this data, as well as the leak incident rates. All of the leaks occurred on pipe which was not capable of passing a smart pig. This pipe was the oldest, with a 1951 mean year of pipe construction. However, it operated at the lowest mean operating temperature 72°F. Although all of the leaks occurred on this pipe, the data for the pipe which has been internally inspected was too limited to yield any meaningful results. 4.13 Seasonal Effects The possibility of incident rate variations throughout the year exist for many causes. For example, heavy winter rains could result in increased external corrosion leaks during the winter. Also, heavy summer construction activity could increase third party damage during this period. In an attempt to evaluate such seasonal variations, the leak data was sorted by month of occurrence. Regulated California Hazardous Liquid Pipelines This data is presented in Table 4-13A for the regulated California hazardous liquid pipelines. Most of the leak causes appeared to have random variations throughout the year. Also, the limited data available for most causes made it difficult to identify any trends. However, the following points were noted: Third party damage from farm equipment did not occur from April through August during the entire 10 year study period. The overall leak incident rate was lowest from April through June. 94

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-12B Incidents by Internal Inspections California Crude Oil Pipelines Under Study Incident Rate Comparison Incidents Per 1,000 Mile YearsSTATh.tIR~~ LInternally InspectedInspectable NotInspectedNot PiggableCause of IncidentNumber RateNumberRate:~ NumberRateExternal Corrosion0 0.0000.0064.27Internal Corrosion0 0.0000.0021.42*3rd Party - Construction0 0.0000.0010.71 3rd Party - Farm Equipment 0 0.00 Total 0 : 0.00 Number of Mile Years 24 Percentage of Total Mile Years 2% Total Length Miles 8 Percentage Total Length 2% Mean Year Pipe Constructed ~. 1985Mean Operating Temperature 19 80 Mean Diameter Inches 6.0 Incident Rate Incidents Per 1 0 : 0 59 4% 20 ~. 4% 1972 81 18.8Comparison,000 Mile Years0.000.00 110140494% :46894%1951727.00.717.12 . ...*8642External CorrosionInternal Corrosion*3rd Party - Construction3rd Party - Farm Equipment0Internally Inspected Inspectable Not Not PiggableInspected95

STATE EMASRAI. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-13A Incident Rates By Month of Year Regulated California Hazardous Liquid Pipelines Incidents Per 1000 Mile Years Causeot Incident l. Jan H ..Febi Mar:t:Apr *:May i:Jun Jut. H Aug Se~:l.~Oct.:.i...NÛv~.LDec. ExternalCorrosion 3.65; 3.49 4.98 3.15 3.15 3.82 5.64 3.15: 3.65 3*49 5.97 6.31 Internalcorrosion 0.33: 0.17 0.17 0.17 0.00. 0.33 0.33 0.17 0.00 0.33 0.33: 0.00 3rd Party-Construction 0.50. 0.83 0.50 1.00 0.50 0.50 1.66 0.83. 0.83. 1.66 1.16 0.833rd Party. Farm Equipment 0.66: 0.33 0.33. 0.00 0.00 0.00 0.00 0.00: 0.17 1.00 0.33 0.173rdParty-TrainDerailment 0.00; 0.00: 0.00: 0.00 0.17 0.00 0.00: 0.00! 0.00 0.00 0.00: 0.17 3rdParty-ExtemalCorr o.oo: 0.00. 0.17 0.17 0.00 0.17 0.00 ÿj7~ o.oo: 0.17 0.17! 0.17 3rdParty-Other 0.17: 0.33H 083! 0.17 0.00 0.00 0.17: 0.00 0.00: 0.17 0.00 0.50 Human Operating Error 0.00; o.oo 0.17! 0.00: 0.00: 0.50 0.00: 0.17: 0.00' 0.17 0.33;. 0.00 Design Flaw : 0.00: 0.00: 0.00! 0.00 0.00 0.00 0.00: 0.17 0.00~ 0.00; .0.17 0.00 EquipmentMalfunction 0.33: 0.83 0.33~ 0.00 0.33 0.17 0.33; i.oo: 0.17' Q*33 0.17. 0.50 Maintenance ! 0.00: 0.17 0.00 0.00 0.00 0.00 0.00 0.17: 0.17 0.17 0.17 0.00 Weld Failure 0.33 0.66 0.33. 0.33: 0.00 0.17 0.17 0.33. 0.17~ ÿ*33: 0.17: 0.17 Other I 0.33 0.50: 017! Q.5Q 0.17 0.00 0.50 0.50' 0.33: 0.50 0.50 0.33 Total 6.31 7.30 7.97 5.48 4.32 5.64 8.80 6.64. 5.48 8.30: 9.46! 9.13 15 10 5 0 Incident Rates By Month of Year Incidents Per 1,000 Mile Years- Extemal CorrosionThird Party AllOperating ErrorWeld FailureEquipment MalfunctionOtherJan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Deco~

California State Fire Marshal ~July 1996 ~Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~nRE~~AL California Crude Oil Pipelines Under Study The incident rates by month of year for California's crude oil pipelines under study are shown in Table 4- 13B. Although this data set is far too limited to draw any meaningful conclusion, we noted that none of the external corrosion caused leaks occurred during the dry summer months May through August. 4.14 Pipeline Components Table 4-14 presents a break-down of the pipeline material, sorted by cause, for each incident which occurred on regulated California hazardous liquid pipelines. As noted, nearly 87% of all incidentS occurred in the pipe body itself. Valves were responsible for another 3.1% of the incidents. 2% were caused by longitudinal weld seam failures in the pipe body. 1.6% were caused by failure at welded fittings. The remaining 6.7% were from various other causes. 100% of the incidents from California's crude oil pipelines included in this study spilled from the pipe body. 4.15 Hydrostatic Testing Interval This section presents the leak incident rates for pipelines grouped with various hydrostatic testing intervals. Regulated California Hazardous Liquid Pipelines The hydrostatic testing requirements for regulated California intrastate and interstate pipelines vary significantly. Basically, the regulations for intrastate lines require periodic hydrostatic testing while those for interstate lines require only initial hydrostatic testing. Specifically, the California Government Code ß51013.5 requires hydrostatic testing of intrastate pipelines as follows: Every newly constructed pipeline, existing pipeline, or part of a pipeline system that has been relocated or replaced, and every pipeline that transports a hazardous liquid substance or highly volatile liquid substance, must be tested in accordance with 49 CFR 195, Subpart E. Every pipeline not provided with properly sized automatic pressure relief devices or properly designed pressure limiting devices must be hydrostatically tested annually. 97

STATEFIRE MA~FIAL California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-13B Incident Rates By Month of YearCalifornia Crude Oil Pipelines Under Study Incidents Per 1,000 Mile YearsCauseof IncidentJan Feb Mar! Apr ~May Jun Jul lAug Sep Oct.j~Nov..~PecExternal Corrosion1.62 `0.000.00 0.81 .0.00 0.00 0.00 0.00 0.81 0.00 1.62 `0.00Internal Corrosion0.000.810.00 0.00 i 0.00 0.00 `0.00 0.00 0.00 0.00 0.00 0.813id Party-Construction 0.00`0.000.81 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.003rd Party-Farm Equipment 0.000.00 0.00 0.00 0.00 :0.81 0.00 0.00 `0.00 0.00 0.00 0.00Total 1.620.81 0.81 0.81 :o.oo :0.81 0.00 0.00 0.81 0.00 1.62 0.81 2 1.5 1 0.5 0 Incident Rates By Month of Year Incidents Per 1,000 Mile Years- External CorrosionInternal Corrosion3rd Party - Construction3rd Party - Farm EquipmentJan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec98

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines4 Table 4-14 Incidents by Item Which Leaked, by Cause Regulated California Hazardous Liquid Pipelines PipeLeakCause No.%,1Valve Pump :We~~ Fitting1 Long. WeldNo.'!%l!No.i%i:No.i%No.i%iExternal Corrosion29867.300.00' 0.00 0.0!00.0Internal Corrosion143.200.0.0. 0.00' 0.0~00.03rdParty-Construction62.14.0212.5~00.01! 12.5~00.03rdParty-Farm'Equipment3rdParty-TraiÒDerailment3rdParty-ExtemalCorrosion18274.10.5,1.60000.0~0.0~0.001 0.0~Dl 0.0!0 0.0k0~ 0.0~0~ 0.0~:0! 0.Oli000.0.00.00.0 3rdParty-OtherHuman Operating Error1352.91.1110!1'0.0~6.3;!O~ 0.011DI 0.0~ 11 12.5~ 1 12.5~Dlo;0.00.0Design Flaw0.0.016.3~JOl 0.0~ii 12.5I~0~0.0Equipment Malfunction 5 Maintenance . 1 Weld Failure 4 OIh~r 13 Total 443 1.4! O.2~ 0.911 2.91100.015:30;4!16'3l.3l~18.811 O.0~25,011100.O~2: 40.0110! 0.D~01 o.oil31 60.0~[51 100.01!01 0.0~Di o.oil4~ 50.0!~01 0.0~81 ioo.o;f 1: 01 8~ 110110.0 0.080.010.0ioo.o `..,irm vveid Threao bonn ~one~ ~onn Other LeakCause ExtemalCorrosion 2 100.0! 0 0.0. 0 0.0 4~ 17.4 Internal Corrosion 0. 0.0 Oj 0.0 01 0.0 0 0.0 3rd Party. Construction 0; 0.0' 01 0.0 0 0.0 0 0.0 3rdParty-FarmEguipment 0 0.01! 01 0.0' 01 0.OH 0 0.0 3rdParty-TrainDerailmerit 0 0.0!l UI 0.0 DI 0.0~l 01 0.0 3rdParty-ExtematCorrosion 0 0.0!! 01 0.0 Dl 0.01~ 0 0.0 3rd Party - Other 1 0. 0.011 0! 0.0:, Dl o.oll 0 0.0 Human Operating Error 0 0.011 Dl 0.01! 0~ 0.0il 1 4.3 Design Flaw 0 0.0 0; 0.O~ Di O.OI~ 0 0.0 Equipment Malfunction 0! 0.011 01 0.011 0: 0.011 13 56.5 Maintenance 0 0.011 Dl 0.011 1 25.0:1 ol 0.0 Weld Failure 0, 0.011 Oi 0.0!: 0! 0.0!! 3. 13.0 Other ii 0! 0.011 oi o.01 ` 3 75.0!! 2! 8.7 Total 1! 21 bOor 0' 00: 4! ioo.oil 23! 100.0 100% - 86.7% 80%- 60% - 40% 20%- 3.1% 1.0% 1.6% 2.0% 0.4% 0.0% 0.8%0%- Pipe Pump Long. Weld Thrd. Conn. Other Valve Welded Fitting Girth Weld Bolted Conn.99

California State Fire Marshal"~ J~IJ July 1996STAThRgEMAKS~iAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines * Every pipeline over 10 years of age and not provided with effective cathodic protection must be hydrostatically tested every three years, except for those on the State Fire Marshal's list of higher risk pipelines which shall be hydrostatically tested annually. * Every pipeline over 10 years of age and provided with effective cathodic protection shall be hydrostatically tested every five years, except for those on the State Fire Marshal's list of higher risk pipelines which shall be tested every two years. * Piping within a refined products bulk loading facility shall be tested every five years for those pipelines with effective cathodic protection and every three years for those pipelines without effective cathodic protection. For interstate pipelines, 49 CFR 195.300 requires hydrostatic testing of newly constructed pipelines; existing steel pipeline systems that are relocated, replaced, or otherwise changed; and onshore steel interstate pipelines constructed before January 8, 1971, that transport highly volatile liquids. The data was reviewed to evaluate hydrostatic testing effectiveness. Two separate pieces of information were gathered. First, the total number of hydrostatic tests performed on each pipeline during the ten year study period was gathered. Secondly, for each leak which occurred during the study period, the date of the preceding hydrostatic test was obtained. To determine the average hydrostatic test interval for each pipeline during the study period, the ten year study period was divided by the total number of hydrostatic tests performed during the study period. Incident rates were then determined for each pipeline within given ranges of hydrostatic testing intervals. Table 4-1 5A presents the resulting data. As indicated, the pipelines which were hydrostatically tested most frequently, up to two years average hydrostatic test interval, suffered the highest leak incident rate. However, these lines were the oldest, operated at the highest mean operating temperature, and had the smallest mean diameter. All of these factors would tend to increase the incident rate. On the other end of the spectrum, the lines which had the longest average hydrostatic test interval suffered the lowest leak incident rates. But these lines were the newest and had the lowest mean operating temperature. Once again, these factors would tend to decrease their incident rates as we have already seen. California's higher risk pipeline category would also tend to skew this data. As previously mentioned, these lines had a generally much higher leak incident rate. Those which were greater than 10 years old were required to be tested at either one or two year intervals, depending on whether or not they were cathodically protected. 100

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines srAl~~,~utru. Table 4-1 5B presents the second set of data - the time since hydrostatic testing for each leak, regardless of cause. Although not as drastic, this analysis resulted in similar results. As indicated, the pipelines which had the shortest interval between hydrostatic testing and the leak, suffered the highest leak incident rate. However, these lines were the oldest, operated at the highest mean operating temperature, and had the smallest mean diameter. All of these factors would tend to increase incident rates. On the other hand, the lines which had the greatest length of time between hydrostatic testing and the subsequent leak, had the lowest leak incident rates. But these lines were the newest and had the lowest mean operating temperature. As has been previously noted, these factors would tend to decrease their incident rates. With the data presented, it is difficult to readily determine the effectiveness of hydrostatic testing. The multiple regressions indicated that pipe age and operating temperatures had the greatest impact on leak incident rates. We believe that the data presented in this subsection reflected the pipe age and operating temperature effects. From these data alone, it is impossible to determine whether or not more frequent hydrostatic testing affected the frequency of leak incidents. However, using these data, we do not conclude that more frequent hydrostatic testing reduced leak incident rates. California Crude Oil Pipelines Under Study In contrast to the regulated California hazardous liquid pipelines, there are no requirements for hydrostatic testing California's crude oil pipelines under study. As a result, it was not surprising to find that 87% of these lines had never been hydrostatically tested; 2% had been tested within the last five to ten years; 1% had been tested within the last two to five years; and 10% had been tested within the last two years. The distribution and incident rates for these crude oil pipelines is presented in Table 4-l5C. Unfortunately, there was insufficient data to allow a meaningful analysis. However, the vast majority of the lines, which had never been tested, had a leak incident rate of 5.42 incidents per 1,000 mile years. This value is less than the 6.72 incidents per 1,000 mile year incident rate for all of the pipelines included in this study. Further, the leak incident rate for external corrosion caused leaks was 3.10 incidents per 1,000 mile years, versus 4.03 incidents per 1,000 mile years for all of the crude oil pipelines under study. The highest incident rate occurred on the pipelines which had been tested within the last two to five years. This group suffered a leak incident fate of 167 incidents per 1,000 mile years. However, this resulted from only two leaks on about four miles of pipelines. 101

California State Fire MarshalJuly 1996srAT~c~E ~t~isn~ Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-15A Average Hydrostatic Testing Interval During Study Period Regulated California Hazardous Liquid Pipelines Incident Rate Comparison Incidents Per 1,000 Mile Years Up to 2.0 Years :~ 2.1 to 5.0 Years :~ *,..: ~ .iaOYears ::; cause of Incident Total No. Rate Number 4~I Rate.! ~ External Corrosion 144 9.58 1i3 4.67: 36 2.06 Internal Corrosion 6 0.40 6 0.25 0 0.00 3rd Party-Construction 21 1.40 15. 0.62: 16. 0.92 3rd Party Farm:Eguipment Oi 0.00 61 0.25! 11 0.63 3rd Party - Train Derailment 0! 0.00~ 01 0.00' 1 0.06 3rdParty-ExternalCorrosion 2: 0.13;: 4 0.17' 0 0.00 3rd Party:-Other 5 0.331! 2! 0.081 0! 0.00 Human Operating Error 5 0.33~ 3: 0.12! Ci 0.00 Design Flaw 0! O.U0; 1 0.04! 0.00 EguipmentMalfunction 12! 0.8011 9 0.37! 4! 0.23 Maintenance 0 O.OO~ 3! 0.12! ol 0.00 WeldFailure 3: 0.2Cr 10! 0.41! 21 0.11-- Other 3 0.20~! 12: 0.501 41 0.23 Total 201. 13.37!! 184. _________________ NuthbË~bf:MilÈYËai~ ..::~ 15,032~.~.;:~Mean..YearPipe.:COnstructed.. 1949~~~Mean Operating!Temperatu ~:~Fj. ~..Mean DiamÈterflÒËhes ~ I - ! ! I 7.61! 74' 24.173~~~ 17.44 1953~~~ 1~ ~ 88.5~ 12.7~... 12.3~ - External Corrosion ~ Third Party All = Operating Error ~ Weld Failure ~ Equipment Malfunction ~ OtherIncident Rate Comparison Incidents Per 1,000 Mile Years15 ~.105~ V j7~5.1 to 10.0 YearsUp to 2.0 Years2.1 to 5.0 Years

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-15B Time Since Last Hydrostatic Test At Time of Leak Regulated California Hazadous Liquid Pipelines Incident Rate Comparison Incidents Per 1,000 Mile YearsSTAThFIKE Up to 2.0 Years I..2.1 to 5.0 Years 5.1 to 10.0 Years Cause of Incident Total No. I Rate Number Rate Number I Rate TotalNumberofMileYears114.953.Mean Year Pipe Constructed1 94~'Mean Operating Temperature "F1 22.Mean Diameter Inches Time Since Last Hydrostatic Test At Time of Leak Incident Rate Comparison - All Causes Incidents Per 1,000 Mile Years lime Since Last Hydrostatic Test At Time of Leak15Up to 2.0 Years 2.1 to 5.0 Years 5.1 to 10.0 Years103

STATEnSE MARSHAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-15C Hydrostatic Testing Interval California Crude Oil Pipelines Under Study Incident Rate Comparison Incidents Per 1,000 Mile YearsUp to 2.0 Years 2.1 to 5.0 Years 5.1 to 10.0 YearsNoneCause of Incident Number Rate Number Rate Number Rate.Number RateExternal Corros~n 0 0.00 2 166.6700.0043.10 -Internal Corrosion ~, 0 0.00 0 00000.0021.553rd Party -Construction 0 000 0 0.00 1 L 28.5700.003rd Party - Farm Equipment 0 0.00 0 0.0000.0010.77 -. Total 0 -- 0.00 2 1 20 I F 7 5.4 - Numberof Mile Years 149 12 1 1291 Mean Year Pipe Constructed 1966 1951 1989 1950Mean Operating Temperature °F 80 68 141 62 Mean Oiameterlnches I 11.0 8.0 11.0 6.9* External CorrosionI ~. 3rd Party - Construction Internal Corrosion~ 3rd Party - Farm EquipmentIncident Rate Comparison Incidents Per 1,000 Mile Years200150100~50 0Up to 2.0 Years 2.1 to 5.0 Years 5.1 to 10.0 Years None1 fl.t

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Based on this data, hydrostatic testing does not appear to categorically result in a reduction in the leak incident rate. The reader should note that these data may be misleading. Often, operators hydrostatically test lines with a relatively high history of leaks as a preventive maintenance measure. In this way, they attempt to identify the weak points in the pipeline. When a leak develops during the hydrostatic test, water is spilled instead of oil. This prevents significant environmental damage and allows the operator to repair or replace a damaged section of pipeline and prevent crude oil spills. As a result of this practice, the leak incident rates for frequently tested pipelines would be higher, since the lines selected for testing would have a higher incidence of leaks. The results would then indicate that frequently tested pipelines had a higher leak incident rate, while in reality, the hydrostatic tests may have been a very helpful tool for preventing and/or minimizing the number of future leaks. 4.16 Spill Size Distribution In many instances, the spill volume is often related to the amount of environmental and/or property damage involved with an incident. While the first barrel spilled usually causes the greatest damage per barrel spilled, additional spill volume most often tends to increase the environmental and property damage to some degree. This section presents and compares the spill volume distribution data for both regulated California hazardous liquid pipelines and the California crude oil pipelines under study. As is noted, the spill volumes from the crude oil pipelines in this study are much lower than those from the regulated California hazardous liquid pipelines. Regulated California Hazardous Liquid Pipelines The spill size distribution for the regulated California hazardous liquid pipeline leak sample is presented in Tables 4-16A and 4-16B. The coordinates of a few selected points along the curve are summarized below: 27% of the incidents resulted in spill volumes of one barrel or less. The median spill volume was five barrels. 61% of the incidents resulted in spill volumes of 10 barrels or less. 67% of the incidents resulted in spill volumes of 25 barrels or less. 82% of the incidents resulted in spill volumes of 100 barrels or less. * 90% of the incidents resulted in spill volumes of 650 barrels or less. * 95% of the incidents resulted in spill volumes of 1750 barrels or less. The largest spill volume was 31,000 barrels. The large difference between the 5 barrel median spill size and the 408 barrel average spill size was caused by a relatively small number of incidents which resulted in large spill volumes. This increased the average value considerably. 105

STATE PIREM. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 100% 80% 60% 40% Table 4-16A Spill Size Distribution Regulated California Hazardous Liquid Pipelines Spill Size versus Cumulative Percentage of Incidents Spill Size Distribution Spill Size versus Cumulative Percentage of Incidents 010 100 Barrels Only010 20 Thousands Spill Size Barrels3040 20% 0%100% 80% 60% 40% 20% 0%020 40 60 80 Spill Size BarrelsNote: 64.48% of the incidents resulted in spills of 20 barrels or less.1001 n~c

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 200 U, C a, C C 0 I- ~1 .0 E 100 z >` C. C a ~50 a LL 0 150 Table 4-16B Spill Size Distribution Regulated California Hazardous Liquid Pipelines Spill Size Distribution.10 to 0.99 5to9 50to99 1,000to9,999Ito 4 10 to 49 100 to 999 10,000 to 31,000 Spill Size Range Barrels107

California State Fire Marshal July 1996STAThRRE ?~L Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines California Crude Oil Pipelines Under Study Although the data sample is very small, the spill size distribution for California's crude oil pipelines under study are presented in Tables 4-16C and 4-16D. It is recommended that the reader compare Tables 4-1 6A and 4-1 6C, for the larger regulated California hazardous liquid pipelines and the crude oil pipelines under study, respectively. The shape of the curves are nearly identical. However, the x- axis coordinates are smaller for the crude oil pipelines under study. The data points were determined by fitting a logarithmic curve to the limited data points. This spill size distribution data is useful in establishing the likelihood, or return interval, of a given size leak from a given pipeline. By combining the leak incident rate and the spill size distribution data, the probable return interval of various sized spills can be determined. The following leak incident rates for various sized spills were established using these data.Spill Size :* California Crude OilPipelines Under StudyRegulated California*Hazardous Liquid Pipelinesany size leakN/A 7.08 incidents per1.000 mile yearsI barrels or greater 6.72 incidents per1,000 mile years 6.54 incidents per1,000 mile years10 barrels or greater 2.02 incidents per1,000 mile years 3.29 incidents per1,000 mile years100 barrels or greater 1.10 incidents per1,000 mile years 1.42 incidents per1,000 mile years1,000 barrels or greater 0.69 incidents per1,000 mile years 0.58 incidents per1,000 mile years10,000 barrels or greater 0.00 incidents per1,000 mile years 0.075 incidents per1,000 mile years108

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines100%80%60%40%20%0%0 Table 4-16C Spill Size Distribution California Crude Oil Pipelines Under StudySpill Size versus Cumulative Percentage of Incidents"STATfl FIRE1,400 100% 80% 60% 40% 20% 0% Spill Size Distribution Spill Size versus Cumulative Percentage of Incidents 0 to 100 Barrels Only200 400 600 800 1,000 1,200 SpUl Size Barrels 0 5 10 15 20 25 Spill Size BarrelsNote: 80.00% of the incidents resulted in spills of four barrels or less.109

~q. California State Fire Marshal July 1996 STATE MAJCEHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Spill Size Distribution California Crude Oil Pipelines Under Study Incidents Per 1000 Mile Years Spill Size Barrels j No. of Incidents Percentage Cumulative % lto4 1 8 80% 80% 5to9 0 0% 80% 10to49 1 1 10% 90% 50 to 99 0 0% 90% 100to999 0 0% 90% 1000 to 9999 1 10% 100% Total 10 100% 100% Spill Size Distribution 10 U, 0 0 a E z >`4 0 C 05 to 9 10 to 49 50 to 99 100 to 999 1000 to 9999 Spill Size Range Barrels110

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STAkt~MAJ~SKAL As indicated, the incident rate for various sized spills from the crude oil pipelines under study are generally less than those from the regulated California hazardous liquid pipelines. As noted above, the probable return interval from a given length of pipeline can be determined using these data. Often, this data provides a more useful result. This data is presented below for a one-mile pipeline. . .. .Califorma Crude Oil ~Pipelines Under Study : -Return Interval from . any OneMileof.:. * * .: ..FJpehne.::~ . ...Regulated California . Hazardous Liquid . Pipelines - .Return Interval from . any One Mile of . Pipelineleak N/A141 yearsor greater 149 years153 yearsor greater 495 years304 yearsor greater 909 years704 yearsor greater 1,450 years1,720 yearsor greater infinite13,300 years This data can also be analyzed to determine the probable recurrence interval for various sized spills from all of the 7,800 miles of regulated California hazardous liquid pipelines and 496 miles of crude oil pipelines under study. California Crude Oil Pipelines Under StudySize Return Intervaifrom 496 miles of these* PipelinesRegulated CaliforniaHazardous Liquid PipelinesReturn Interval from7,800 miles of these Pipelinesleak N/A 6.6 days, orss lealcs per year 3.6 months, oror greater 3.3 leaks per year 7.2 days, or51 leaks per yearor greater 1.0 YearS 14 days, or26 leaks per yearor greater 1.8 years1.1 months, or11 leaks per yearor greater 2.9 years 2.7 months, or4.5 leaks per yearIll

STATE FIRE MAI~SflAL California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesCalifornia Crude OilPipelines Under StudyRegulated CaliforniaHazardous Liquid PipelinesSpill SizeReturn Interval from 496. miles of these PipelinesReturn Interval from7,800 miles of these Pipelines10,000 barrels or greaterinfinite1.7 yearsAs indicated, because of the relatively small length of crude oil pipelines under studyand the lower frequency of a given size spill, the return interval for a given sized spillfrom these crude oil pipelines is far greater than for the regulated Californiahazardous liquid pipelines. 4.17 Damage Distribution The property damage distribution was very similar to the spill size distribution discussed in the preceding section. A few incidents resulted in relatively large property damage values which increased the mean values considerably. To the greatest extent possible, the damage figures used in this study included all costs associated with the incident e.g. value of spilled fluid, clean-up, injury, judgements, fatalities, etc.. Regulated California Hazardous Liquid Pipelines Table 4-1 7A depicts the property damage distribution data for California's regulated California hazardous liquid pipelines. All data has been shown in constant 1994 U.s. dollars. The values for each year were converted to 1994 constant dollars using the U.S. City Average Consumer Price Indices as published by the U.S. Bureau of Labor Statistics. A few points along the curve are presented below: 25% of the incidents resulted in damages of $2,000 or less. The median damage was $1 1,000 per incident. 75% of the incidents resulted in damages of $57,000 or less. 90% of the incidents resulted in damages of $270,000 or less. 95% of the incidents resulted in damages of $880,000 or less. The largest reported damage for a single incident was $17,500,000. However, we understand that this figure may increase as additional claims are settled.112

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines100%80%60%40% -20%-Table 4-17A Property Damage DistributionRegulated California Hazardous Liquid PipelinesSTAT~ 0 5 Damage Distribution Logarithmic Scale 10,000 Damage $US 1994 100,000,00 Note: 80.72% of the incidents resulted in damage of $100,000 or less. MillionsDamage $US 19941015 0%100%80%60%40%20% 0%10 1,000 100,000 10,000,000 100 1,000,000113

California State Fire Marshal`~ ~ July 1996SrA FIRE MA~HAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines California Crude Oil Pipelines Under Study Although this data sample is very small, the damage distribution data for California's crude oil pipelinesunder study are presented in Tables 4-17B. In comparing Tables 4-1 7A and 4-1 7B, the shape of the curves are nearly identical for the regulated California hazardous liquid and crude oil pipelines in this study, except for the incidents which caused extensive property damage. Although the occurrence of spills which resulted in modest amounts of property damage were essentially the same for both groups of pipelines, the frequency of spills which caused $10,000 $US 1994 in property damage or more was much greater for the regulated California hazardous liquid lines. These data are useful in establishing the likelihood, or return interval, of a leak resulting in a specific amount of damage from a given pipeline. By combining the leak incident rate and the damage distribution data, the probable return interval of various spills for a given pipeline can be determined. The following leak incident rates were established using these data.Damage Resulting From Spill $US 1994Crude Oil Pipelines Under StudyRegulated CaliforniaHazardous Liquid Pipelines I$100 damage 6.72 incidents per1.000 mile years 6.85 incidents per1,000 mile years$1,000 damage 6.72 incidents per1,000 mile years 5.80 incidents per1,000 mile years$10,000 damage 1.34 incidents per1,000 mile years 3.64 incidents per1,000 mile years $100,000 damage. 1.14 incidents per1,000 mile years 1.36 incidents per1,000 mile years$1,000,000 damage 0.00 incidents per1,000 mile years 0.28 incidents per1.000 mile years$1 0.000,000 damage 0.00 incidents per1.000 mile years 0.028 incidents per1,000 mile years114

California State Fire MarshalJuly1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines100%80%60%40%20% Table 4-17B Property Damage DistributionCalifornia Crude Oil Pipelines Under StudySTA fiRE MM~SW. 200 Thousands Damage $US 1994 1 00,000~000 100300 400Damage DistributionLogarithmic Scale 0%100%80%60%40%20% 0%10 1,000 100,000 10,000,000100 10,000 1.000,000 Damage $US 1994115

California State Fire Marshal July 1996STAM~sH.~. Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines As noted in the previous section, the probable return interval from a given length of pipeline can be determined using these data. Often, this provides a more useful result. These data are presented below for a one-mile pipeline.Damage Resulting From Spill $US 1994California Crude OilPipelines Under StudyReturn Interval from any One Mile of PipelineRegulated California Hazardous Liquid Pipelines -Return Interval from any One Mile of pipeline$100 damage149 years146 years$1,000 damage149 years172 years$10,000 damage746 years275 years$100,000 damage1,090 years735 years$1,000,000 damageinfinite3,570 years$10,000,000 damageinfinite35,700 years These data can also be analyzed to determine the probable recurrence interval for various sized spills from all of the 7,800 miles of regulated California hazardous liquid pipelines and 496 miles of crude oil pipelines under study.Spill SizeSUS 1994California Crude OilPipelines Under Study -Return Interval from 496 miles of these PipelinesRegulated CaliforniaHazardous Liquid PipeiinesReturn Interval from7,800 miles of these Pipelines$100 damage 3.6 months, or3.3 leaks per year 1 week, or53 leaks per yearS 1.000 damage 3.6 months, or3.3 leaks per year 7.8 days, or45 leaks per year$10.000 damage1.5 Y~&5 12 days, or28 leaks per year$100,000 damage2.2 years1.1 months, or11 leaks per yearS 1,000,000 damageinfinite 5.5 months, or2.2 leaks per year$10,000,000 damageinfinite4.6 years1 i~c

California State Fire MarshalRisk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines st~ r L As indicated, because of the relatively small length of crude oil pipelines in this study and the lower frequency of spills resulting in relatively large values of damage, the return interval for spills from these pipelines resulting in significant damage is greater than for the regulated California hazardous liquid pipelines. 4.18 Incident Rates by Internal Coating or Lining The possibility for significant internal corrosion was envisioned on the crude oil pipelines evaluated in this study. As a result, data regarding the installation of internal liners or coatings was gathered. However, only about 1% of the pipelines had an internal liner or coating installed. Although not statistically relevant, all of the leaks occurred on unlined or uncoated pipe. The data sample was too small to facilitate an analysis of this parameter. 4.19 Incident Rates by Above versus Below Grade Pipe 96.3% of the 496 miles of crude oil pipelines under study was buried below grade. Of the remaining pipe, 3% 15 miles was installed above grade, and 0.7% 3 miles was installed with a combination of both buried and above grade segments. Table 4-19 presents the incident rates for the above and below grade pipelines. As indicated, all of the leaks occurred on the buried sections of line. However, this should not be considered statistically relevant because of the very limited data sample. 4.20 Recovery of Spilled Volumes Although only 10 leaks occurred during the three year crude oil pipeline study period, the relationship between the volumes spilled and the volume recovered was reviewed. As indicated in Table 4-20, of the 1,221 barrels of crude oil spilled, roughly two- thirds 800 barrels were recovered. The lowest recovery percentage occurred from the external corrosion leaks. This relationship is not surprising, since these leaks are typically very slow, low leak rate incidents. 117

STATE FIRE MA~HAL California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Table 4-19Incident Rates By Above vs. Below GradeCalifornia Crude Oil Pipelines Under Study Incidents Per 1,000 Mile YearsCause of Incident Above Below BothExternal Corrosion 0.00 4.19Internal Corrosion 0.00 1.400.00 --0.003rd Party- Construction 0.00 0.700.003rd Party - Farm Equipment 0.00 0.70- 0.00T0t81 1 0.00 6.980.00Number of Mile Years 45 143210Mean Year Pipe Constructed 1978 19521947Mean Operating Temperature ~F 86.7 74.360Mean Diameter Inches 3.2 7.7 I 5.5Average Spill Barrels 0 122..1 0Average Damage $US 1994 $0.00 $39,020.00 $0.00 8 6 4 2 0- External CorrosionInternal Corrosion~..:S 3rd Party - Construction3rd Party - Farm EquipmentIncident Rate Comparison Incidents Per 1,000 Mile YearsAbove Below Both110

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~nRE~tHAL 4.21 Injuries and Fatalities Regulated California Hazardous Liquid Pipelines 49 injuries and 3 fatalities resulted from incidents on regulated California hazardous liquid pipeline system during the ten year study period. Nearly 94% of the injuries and 100% of the fatalities resulted from only three incidents; it is remarkable that just over one-half percent of the total incidents resulted in all of the fatalities and nearly all of the injuries during the entire ten year study period. These incidents are briefly described below: * May 25. 1989. San Bernardino - On May 12, 1989, a freight train derailed in San Bernardino, California. On May 25, 1989, 13 days later, a regulated interstate petroleum products pipeline ruptured. The National Transportation Safety Board determined that during the derailment, and later during the movement of heavy equipment to remove the wreckage, the high-pressure products pipeline adjacent to the tracks was damaged and weakened. Less than two weeks after the wreck, the pipeline ruptured and spilled more than 300,000 gallons of gasoline into a nearby neighborhood. Some of the gasoline ignited and caused significant fire damage. This incident resulted in two fatalities and thirty-one injuries. February 22. 1986. Placer County - During the removal of an abandoned section of pipeline which had been relocated around a collapsed railroad trestle, approximately one barrel of gasoline was spilled. The fuel was ignited by a torch being used by the railroad's welding crew. As a result of the ignition, three welders jumped from the bridge into the creek below. This incident resulted in one fatality and one injury. November 22. 1986. Tustin - A ten-inch API 5L X52, ERW pipe longitudinal weld seam ruptured. This resulted in the spill of about 11,000 barrels of unleaded gasoline. Fortunately, the spill did not result in fire or an explosion. Documents filed with the USDOT indicated that there were no injuries or fatalities meeting federal reporting criteria. See also Section 3.0 of this study. However, 14 emergency responders from the local fire department were treated for symptoms consistent with hydrocarbon exposure: eight were treated at a medical facility, four were treated and released at the scene, and one was hospitalized for observation. In addition, one civilian was also treated at the scene and released. These were treated as 14 injuries for the purposes of this study. Each of these incidents had a different cause. Two were caused by some form of third party damage, while the third was caused by a material defect. 119

California State Fire Marshal I4~1J July 1996STAT RM~SHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The number of incidents resulting in injuries and fatalities was too small to draw any meaningful conclusions. However, it should be noted that all injuries and fatalities occurred on petroleum product pipelines. Crude line incidents did not result in any injuries or fatalities during the study period. The current requirements are basically the same for product and crude pipelines. However, although a limited sample, this data indicated that the risks to human life were likely greater for product pipelines. On the other hand, both crude and product pipeline incidents resulted in similar environmental concerns. As mentioned previously, all injuries, regardless of severity, were included in these data. For instance, the 1986 Tustin incident resulted in 14 injuries which did not meet the USDOT injury reporting criteria. Deleting these injuries alone would have reduced the resulting injury rate for this study by more than one-third. The reader should keep this factor in mind while reading this section. Otherwise, the public injury risk could be over-exaggerated. Sufficient data was not available to sort the injuries incurred during the study period by severity. California Crude Oil Pipelines Under Study No injuries or fatalities occurred on the California crude oil pipelines during the three year study period. Further, the data sample was too small to be meaningful. For example, if one simply applied the fatality rate or 0.042 fatalities per 1,000 mile years, which resulted from the regulated California hazardous liquid pipelines, one would anticipate a fatality every 16 years for the 496 miles of crude oil pipelines included in this study. This recurrence interval is greater than the 3 year study period. As a result, one would not expect a fatality during this study. Further, as discussed above, the risk to human life from crude oil spills is likely less thanior refined petroleum product pipelines, which would tend to increase the recurrence interval. 120

~e~4California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines5.0 Barriers and Incentive Options As noted earlier in Section 1.0, California Government Code Section 51015.05 requires that the CSFM investigate incentive options that would encourage pipeline replacement or improvements, including, but not limited to, a review of proposed regulatory, permit, and environmental impact report requirements and other proposed public policies that could act as barriers to the replacement or improvement of these pipelines. To this end, on January 31, 1996, EDM Services distributed a questionnaire regarding incentive options and barriers to pipeline replacements and/or improvements. 231 questionnaires were distributed to: operators of regulated California Hazardous Liquid Pipelines, all participants in this study, regulatory and jurisdictional agencies, coastal communities with a high density of oil and gas activity e.g. San Luis Obispo, Santa Barbara, and Ventura, and the Pipeline Assessment Steering Committee. The questionnaire contained 14 questions designed to gather information on, measure attitudes toward, and obtain suggestions about proposed or potential incentives and barriers to pipeline replacement and/or improvement. Respondents were allowed one month to complete the written questionnaire, although considerable latitude was given to those who needed additional time. In all, 28 responses were received; a rate of response well within the bounds of acceptability for this method of study design and implementation. Nine completed questionnaires were obtained from regulatory or jurisdictional agencies and 19 were received from operators both majors and independents. In addition, nine of the respondents stated in, one form or another that their company/agency could offer no comments to the CSFM on these particular issues. One respondent provided comments only on the initial permitting process due to a lack of experience in replacing or improving pipelines. Though not specified in the questionnaire, respondents were allowed to provide answers and case studies for pipeline projects that are not included in AB 3261 or otherwise a part of this study. The responses were analyzed by BDM Oklahoma. This section 5.0, was authored by BDM Oklahoma's Deborah Pratt and Jerry Simmons, using the responses received. In the following analysis of the questionnaire results, some classifications and groupings of answers have been employed. First, a distinction was made between responses from regulatory agencies, on the one hand, and private companies on the other; due largely to observable differences in emphases and in the qualitative nature of the responses. Second, with regard to incentive options, a distinction was made between what can be termed "negative" and "positive" incentives. "Negative" incentives refer to those actions or suggested actions taken by government agencies in response to pipeline leaks, non- 121

ji~ California State Fire Marshal July 1996STATeFU~EMA'1CSNAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines compliance, etc. These incentives are often punitive in nature and seek to deter undesirable behavior or correct it after the fact. "Positive" incentives refer to those actions taken by regulatory agencies that seek to reward operators who have a history of sound regulatory compliance, thus engendering continued attention to issues of pipeline safety. 5.1 Summary of Questionnaire Results: Incentives The first set of questions targeted potential and/or proposed incentive options available to regulating agencies. In each case, respondents were asked to identif~' incentives that would encourage pipeline replacements or improvements and indicate how these incentives should be implemented. The reader should note that respondents were not required to rigidly adhere to the format, but were afforded the opportunity to fully explain their responses and provide case studies where appropriate. The following is a summary of the responses to potential and/or proposed "incentives options". A. Responses from Regulatory Agencies Negative Incentives The most commonly cited potential and/or proposed negative incentives by regulatory agencies pertain to, in one form or another, changing the nature and scope of the consequences of pipeline leaks and non-compliance. Possible consequences included: civil penalties, - require replacement or re-conditioning of sections of pipeline that have "excessive" leak history, require reduced operating pressures for pipelines with "excessive" leak history, * increase inspection of poorly maintained pipelines with identified integrity problems, and * assess all annual fees based on the degree to which the pipeline is "leak prone". Positive Incentives The following "positive" incentives were most often suggested by the regulatory agencies: reduce inspection of new pipelines after a sound regulatory compliance history has been established, 122

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines sr*i~sjw. * extend the time between required hydrostatic tests under state law for new or replaced pipelines, allow operators to use an alternative test method in lieu of the hydrostatic test, * provide a "good service award" for the pipeline company with the most reconditioned or replaced sections of pipeline, * provide assistance financial and otherwise to companies that are obtaining permits and authorizations to do replacements and/or improvements, * adopt "regional guidelines and processes" for pipeline activities that promote environmental, safety, and health concerns, * reward compliant operators with expedited government reviews, * establish cooperative emergency response planning and resources, and * categorically exclude pipeline replacements or improvements from the California Environmental Quality Act CEQA. B. Responses from Pipeline Operators Not surprisingly, suggested incentives by operators were skewed toward the positive side; that is, the number of positive incentives exceeded the number of negative incentives by almost nine to one. The following is a summary of the incentives proffered by the operators: Negative Incentives * fine companies for every leak, incident, or other "negative" situation. Positive Incentives reduce audit frequencies/scope, streamline the permitting process, i.e. "one-stop-shopping", reduce the frequency with which hydrotesting must be conducted, reduce/eliminate CSFM fees on pipelines that have been replaced or improved, 123

c'~~ California State Fire Marshal July 1996 STAT~FIRE Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines * provide for an automatic negative declaration of adverse environmental impact for pipeline replacement or repair projects being done to improve safety, * formally recognize operators and individuals, e.g. positive press releases by CSFM, plaques, letters, notices of commendation, annual luncheon/dinner to recognize pipeline safety achievements, etc., * establish a fund to reimburse or partially reimburse corporate investments in technologies that reduce leaks and incidents, ensure compliance, etc., and establish an Operator Pipeline Safety Leadership Committee to provide ongoing recommendations to CSFM on pipeline safety issues. 5.2 Incentive Implementation There were very few specific responses which provided input regarding how these incentives could be implemented. However, the idea of establishing some sort of a task force garnered support from both regulatory agencies and operators. The following implementation suggestions were offered by the participants. * Install joint industry/government task force in a partnering process to review promising ideas and determine feasible implementation, Implement all incentives at the regulatory "staff' level as opposed to hearings, appeals processes, etc. Leave the decision to replace or repair a pipeline "solely with the individual pipeline operator." 5.3 Summary of Questionnaire Results: Barriers The second set of questions targeted perceived barriers to pipeline replacement projects. Respondents were asked to identify barriers, describe the actual and potential consequences of these barriers, and suggest ways in which the barriers could be mitigated. Although the questionnaire clearly distinguished between barriers and incentives, there was some overlap in the responses to each; that is, similar responses were received for both types of questions. In addition, seven of the nine regulatory agencies did not respond to questions on barriers citing, for the most part, a lack of relevant case histories of projects which have been delayed, deferred or canceled because of regulatory, permit or environmental impact barriers. A significant portion of the responses summarized below, therefore, came from the operators that responded to the questionnaire. 124

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines srAkt.!~MAxSHAL Regulatory Barriers. Permitting Barriers, and EIRs By far the most commonly cited barriers to replacing or improving pipelines involve the permitting process. Operators across the board indicate that these processes: 1 take far too long; 2 demand an unrealistic allocation of expenditures; and 3 may unnecessarily put the environment and the safety and health of the public at risk. Some of the difficulties expressed by respondents include: obtaining construction permits from various cities in a timely manner, obtaining Negative Declaration Status often taking up to 18 months, acquiring an "Endangered Species Management Agreement" 2081 permit, complying with CEQA requirements due to implementation variances from county to county, erroneous application by the Los Angeles Fire Department of city regulations to jet fuel pipelines; or the "more is better" school of regulation, slow responses by Los Angeles Transportation and Public Works Departments One operator stated that it can take up to six months for the Los Angeles DOT to decide on a relevant CEQA standard., California Government Code sections 51013, 51014 `regarding hydrostatic testing, excessive details required by CSFM, franchise agreements, local agency street opening excavation or building permit process, California Coastal Commission and BCDC permit processes, and Environmental Impact Reports. 125

c*"~4. Calilornia State Fire Marshal July 1996 STATP,!~E Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines 5.4 Actual and Potential Consequences of Barriers According to respondents, the actual and potential consequences of the identified barriers are predominately financial; although environmental, safety, and health consequences were also noted with some regularity. There is a tremendous amount of concern among the operators that pipeline improvements/replacements have become so costly and cumbersome that they no longer have any incentive to be proactive in these matters. In fact, one respondent stated that replacements and improvements are now considered "...only as a last resort to all other options." Environment, safety, and health consequences were also noted by some respondents. For example, in one case, an operator proposed to install and operate internal corrosion inhibitor storage and injection facilities at its pump station facilities in a particular County. The initially proposed project took more than 18 months from application submittal to receipt of construction approvals and permits. Although other temporary measures were taken by the pipeline operator, these measures involved more risk that the actual proposed project and delayed the implementation of a more desirable corrosion inhibitor program; fortunately, pipeline integrity was not impacted by this delay. Other commonly cited consequences include: unnecessary and unrealistic expenditures of time and financial resources, * project delays, deferrals, or elimination, * actual amount of pipe replaced is decreased, hydrostatic testing requirements are damaging pipelines, accelerating leaks, and leading to the generation of contaminated waste water, and * marginal gathering lines are no longer being replaced by some operators. 5.5 Removing Barriers The overwhelming consensus of the study's participants is that the permitting process must be streamlined. One of the primary areas of concern involves jurisdictional issues. Many respondents both regulators and operators expressed a desire to eliminate overlapping agency and redundant requirements. As one operator stated, "Our safety and emergency procedures went through at least four iterations before receiving agency approval. A primary reason for these iterations was conflicting comments received from different agencies. It was extremely difficult to satis1~' everyone due to intra- agency jurisdictional disputes." I 2f~

California State Fire Marshal ,, j~July 1996 ~Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~~Z~SHAL Although the respondents consensus was that the permitting process must be streamlined, it should be noted that some local agencies have made recent improvements to improve their processes. One County cited the issuance of minor use permits, instead of the more typical conditional use permits which require Planning Commission approval, for pipeline upgrade projects. Emergency permit processes have also be developed to allow immediate pipeline work when circumstances warrant. With respect to the CSFM in particular, respondents appear to want mechanisms to ensure that counties or other agencies such as the Los Angeles Fire Department do not impose requirements or regulate pipeline safety issues that fall under the exclusive authority of the CSFM. The most common suggestions for jurisdictional streamlining are as follows: * develop Memoranda of Understanding which begin to address problem areas and identify primary agency responsibilities Two respondents cited jurisdictional problems between the CSFM and the Los Angeles Fire Department., * create or designate a single State agency with sole jurisdiction over pipeline issues, and * establish the USDOT as lead permitting agency for pipeline maintenance projects within the Interstate Pipeline Rights of Way. Respondents also provided the following suggestions about specific regulations, possible modifications, exemptions, and timing: consistently implement the Long Term Programmatic Permit for Threatened and Endangered Species among the different BLM Resource Area Offices for maintenance projects, develop a clear procedure or flow chart of required documents, set time limits for BLM to complete permit applications once received by the appropriate office, apply smart pigging requirement to new pipelines only, limit the requirement to upgrade all components within a line section when only a small replacement is required, eliminate periodic hydrostatic testing requirements on existing pipelines, provide categorical exemption under CEQA for pipeline replacement projects under the jurisdiction of the CSFM, 127

~ :~n~g. California State Fire Marshal July1996 STATh~R~MA~SHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines eliminate the county billing method, and exempt pipeline safety replacement projects from EIRs. 5.6 Case Studies Following are a few case studies and excerpts from the completed questionnaires. It should be noted that due to time constraints, many of the local agencies did not have an opportunity to develop specific case study responses. The reader should also note that the information presented in these case studies has not been independently verified, nor conducted a methodologically sophisticated analysis of the results. Hence, the excerpts below should not necessarily be taken as fact or considered to be representative of the entire sample of respondents. The intent of the questionnaire, and of this portion of the report, was to identify public policies that could act, or be perceived as barriers to the replacement or improvement of pipelines and, similarly, to identify possible incentive options that would encourage these activities. The reader should note that the actual responses have been edited to the actual company and agency names. A. "We are attempting to replace and relocate a portion of an acetylene welded pipeline within the City A. A section of this pipeline runs through a school property. The only alternative to relocate this section would be to obtain new right-of-way through the City B. The City B is not cooperating and is essentially telling us that they do not want to take on City A's problem. This delay has caused the pipe not to be relocated to an area safer to the public." B. "CEQA is the most significant regulatory barrier. The implementation of CEQA varies significantly from county to county. Some counties have planning departments that take the CEQA issues to the "nth" degree. As a illustration, a permit from County A for one pump station and one 10.5 mile pipeline has 109 permit conditions ... The permit costs are substantial; the 1995 permit fees from County A for this permit were about $192M. Probably 1/3 of that was attributed to new construction in the pump station. The construction work required a Supplemental EIR that cost in excess of $1 OOM and took over 2 years to get approved ... Most of the pipeline replacement work that we undertake is due to corroded pipe identified from internal inspections smart pigging. We believe that permitting times of 1-2 years is an excessive amount of time to wait when you know that the pipe is corroded. The actual consequences of the permitting barriers is that we do not replace pipe as quickly as we would without the barriers and the amount that we replace is less than it would be if the resource burdens of permitting were less. This tends to increase risk. Also we have stopped replacing marginal gathering lines. The economics of these pipelines can not justify the cost of preparing a development plan or a minor use permit and the expensive permitting process. We have begun to petition the CPUC to begin shutting down these lines. The oil from the leases that these lines serve will have to be trucked." 128

Cahfornia State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines SThTBFIRE r C. "This project was voluntarily proposed to reduce the risk of an environmental incident. The State Lands Commission strongly supported getting this work done but it was strictly up to us to take the initiative to get the permits ... Platform A lies about 2.5 miles off the California coast. It produces about 4,500 barrels of oil per day and 3 million cubic feet of gas per day. The oil and associated water are piped to a separation plant on the beach through a single 6 inch subsea pipeline. The sour gas is piped to the same plant, for sweetening, dehydration and compression, through a separate 6 inch subsea pipeline. The platform was installed along with the two pipelines in 1966. Mitered bends of 30° were used in the pipelines, at the beach, in the surf zone, for a direction change. Miter bends are not typically used for this purpose. Curved or manufactured bends are usually used for direction changes in pipelines. Today, electronic inspection tools known as "smart pigs" are pumped through pipelines to inspect the condition of the pipelines. These tools are usually about 10 to 12 feet long and are segmented to go around bends. The segments are too long to make it through a miter bend, so the miters must be replaced with curved bend pieces if an inspection is to be done on the lines. We would like to electronically inspect the condition of these 1966 vintage pipelines to insure that they are still in good condition. Annual pressure tests of these lines have not resulted in any problems or failures to date. A break or leak in the oil line would of course result in oil getting in the ocean. Replacement of these lines in their entirety would cost 3 to 5 million dollars and would take 2 to 3 years to permit, if permittable at all, under the current permit conditions. This project involves simply cutting out the miter bends and welding in long radius bends. This is essentially four, 6 inch pipeline cuts and eight, 6 inch pipe welds. The previous 50% owner and operator of the Eliwood operation started getting proposals to replace these miter bends in 1983. When we took over operating and 100% ownership in 1993, the previous operator still did not have permits to do thisjob. We started working on a design and permit application in the 2nd quarter. of 1994. This included many meetings with the County staff, the County Fire Department and the County Building and Safety Department to insure compliance with all regulations and to negotiate the conditions imposed by these agencies. The application for County Planning and Development, Energy Division, Development Plan Permit and Conditional Use Permit were officially submitted in December 1994. Additional permits required were: County Coastal Development Plan permit, California Coastal Commission Coastal Development Plan permit, U.S. Army Corps permit which require California Regional Water Quality Control Board RWQCB waiver or certification under the section 401 of the Clean Water Act, and California State Lands Commission approval. 129

California State Fire Marshal July 1996grA~u~I~sBAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines The Coastal Commission, Coastal Development Plan permit is essentially the same as the County CDP but it can't be applied for until the County DP and CUP are approved. The Army Corps approval can't be obtained until the California Coastal Commission CDP is received. The whole process is burdened with redundancy. The County and Coastal Commission look at exactly the same issues and scrutinize these issues independently, wasting large amounts of time and money. The staff report by the County was over 100 pages. The State Lands Commission is the only agency that has the technical expertise to look at the mechanics of how this tidal zone job is being done, yet the procedure was reviewed and scrutinized, and required justification by the County Staff, the Fire Department Staff, the County Building and Safety Staff and the California Coastal Commission. The County invoices for county staff time for this project amount to somewhere in the neighborhood of $40,000 for review and staff report preparation. Remember this is for 4-6" pipeline cuts and 8-6" pipeline welds. Prior to ever getting to the Planning Commission Hearing Board, the County staff places conditions in the staff report on the project and the applicant and County staff have a one sided negotiation on these issues. We had very little leverage to get anything changed. The County takes the opportunity to add operational conditions that have not been required or necessary in the past 30 years of operation and require acceptance in order to get the staff report finalized for the commissioners. One example is, on very rare occasions the beach section of the pipelines become completely uncovered by natural sand transport during the stormy season. Usually this occurs between January and March. We, as prudent operators, always watched the lines to insure they were not damaged or did not move around too much in the surf during time period. A new condition for the remaining life of the pipelines states that we must shut down the entire production operation when more than 20 feet of the 16,000 foot long pipeline is exposed in the surf zone and there are 12 foot high waves. This means we would be required by permit to shut down the production operation under the stated conditions event if there was no risk to pipelines. Another condition is that we must visually inspect the pipeline every day of the year and keep a written log for County inspection. This requirement disregards that over 300 days a year there is absolutely no sign of pipelines on the beach, so this requirement is an expensive waste of manpower. Another extreme condition requires draining the flush water, which is ocean water, from the pipe prior to cutting the pipe. This is following flushing the lines to a point where the flush water had less than 30 ppm Oil and Grease content. To drain the water we will have to hot tap a weld-o-let on the pipeline and drain the flush water out of the section of the pipe uphill of the cut point. This was proposed by us in an effort to get around having a Clean Seas vessel on location and avoid a job shutdown because of a sheen. In addition we are required to have over 400 foot of absorbent boom on site for spill protection. All this for .03 gallons calculated at 30 ppm of oil in the 1000 feet of 6" pipeline which was uphill 130

California State Fire Marshal c¯lO~.July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines of the cut. To summarize, this is a project that we voluntarily took on to protect the environment and insure we would not have an oil spill incident. It is what should be a simple job but will probably end up costing over $250M to do plus 4 days of lost production at $30,000 to $40,000 per day lost revenue. In a normal setting this job would be much less costly and time consuming. It would have been done years ago and there would be many electronic inspection records by now that could be used to develop trends on pipe degradation. We would be able to accurately predict when and if a pipe failure would occur. The economic conditions for the Ellwood Asset have changed recently. We no longer intend to perform this repair until we have determined the future of this operation. The subject of the miter bend replacement would not be at issue now if the permitting process would have been reasonable and timely. The miter bends would have been replaced by the prior operator years ago or by us in 1994. This situation could easily be improved by making one agency responsible for reviewing this type of work. Then have policies that allow practical common sense judgments on issues of how to do the job based on the end result being much better than the current condition. Eliminate the redundancy of multiple agencies looking at the same thing and rely on the agency that has the most technical expertise to review the project. Eliminate the county billing method that encourages 100 page documents for what would be a half day job in another location. There is no incentive for County staff to be efficient and effective." 131

California State Fire Marshal ` July 1996STATE FIRE MAREHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines6.0 Conclusions The conclusions which can be drawn from this study have been organized into two groups. * The first includes those which can be drawn from the pipeline and leak database developed and the study conducted in accordance with California Government Code Section 51015.05. * The second includes those conclusions which can be drawn from the incentive option investigation, also conducted in accordance with California Government Code Section 51015.05. 6.1 Database and Study Findings The following pipelines were included in the data base and study: * pipelines for the transportation of crude oil that operate at gravity or at a stress level of 20% or less of the specified minimum yield strength of the pipe; and, * pipelines for the transportation of petroleum crude oil in onshore gathering lines located in rural areas. Pipelines meeting this criteria have been included in the study and database, whether they were operating or not during the study period; even abandoned, idle, or otherwise out of service pipelines have been included in the study and database. The following pipelines were excluded from the data base and study: interstate and intrastate pipelines which are currently regulated by the CSFM or the USDOT; gathering lines located entirely within the boundary of a DOGGR oil field boundary, or which cross a boundary where two DOGGR oil fields are contiguous and are contained entirely within multiple DOGGR oiifields; flow lines located entirely within the boundary of a D000R designated oil field boundary, or which cross a boundary where two DOGGR oilfields are contiguous and are contained entirely within multiple DOGGR oilfields; natural gas pipelines; refined petroleum product pipelines; and abandoned pipelines which have been physically removed. 132

California State Fire Marshal I~SLJu'y 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines ST*1M~1tS~iAL The database and study analyze California's crude oil gathering pipeline risks, utilizing leak incident data for a three year period, from January 1993 through December 1995. Although extensive efforts were taken to gather the most complete database possible, including the distribution of over 1,200 questionnaires aimed at identifying study participants, the resulting data set was relatively small. The data set can be summarized as follows: Number of Leaks of one barrel or greater - ten 10 * Number of Pipelines - 113 * Total Length of Pipelines - 496 miles * Mean Pipe Diameter - 7.5 inches * Mean Operating Temperature - 74.2°F Cathodically Protected Pipe - 317 miles 64% of total * Bare Pipe - 87 miles 18% of total Median Spill Size - 3 barrels Average Spill Size - 122 barrels Median Damage - $5,000 $US 1994 * Average Damage - $39,020 $US 1994 Percentage of Below Grade Pipe - 96.3% 478 miles Overall Incident Rates The overall leak incident rate for leaks of one barrel or more from the crude oil pipelines under study was very similar to the regulated California hazardous liquid pipelines - 6.72 versus 6.54 incidents per 1,000 mile years respectively. However, the incident rate for larger spills was generally much less for the smaller, crude oil pipelines in this study. The results for these crude oil gathering lines are summarized shown on the following page: 133

TATRFIR~ MARS~4AC. California State Fire Marshal July 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering LinesEvent1 barrel or greater spill10 barrels or greater spillIncident Rate6.72 incidents per 1,000 mile years2.02 incidents per 1,000 mile years100 barrels or greater spill1.10 incidents per 1,000 mile years1.000 barrels or greater spill0.69 incidents per 1,000 mile years10,000 barrels or greater spill0.00 incidents per 1,000 mile years$1,000 damage resulting from spill $US 19946.72 incidents per 1,000 mile years$10,000 damage resulting from spill SUS 19941.34 incidents per 1,000 mile years$100,000 damage resulting from spill $US 19941.14 incidents per 1,000 mile years$1,000,000 damage resulting from spill SUS 19940.00 incidents per 1,000 mile yearsinjury resulting from spill0.00 incidents per 1,000 mile yearsfatality resulting from spill0.00 incident per 1.000 mile years The above spill size and damage values were estimated by connecting the few data points with logarithmic curves. The resulting data are depicted graphically in Tables 6-lA and 6-lB. For comparison purposes, this table also shows the regulated California hazardous liquid pipeline data. As indicated, the smaller crude oil pipelines under study had lower leak incident rates for a larger spill size or damage values. Also, these crude oil lines had maximum spill size and damage figures much less then the regulated California hazardous liquid pipelines. External Corrosion External corrosion was by far the leading cause of incidents, representing 60% of the total. However, with the limited data sample, we were unable to isolate the cause. The results of the 1993 study regarding regulated California hazardous liquid pipelines indicated that pipe operating temperature and age were the two leading factors contributing to increased external corrosion. We suspect that this is also the case for the crude oil pipelines under study. However, the data set was too small to perform a conclusive analysis. Recovery of Spilled Volumes The operators reported that the 10 leaks which occurred during the three year study period resulted in an estimated 1,221 barrels of spilled crude oil. Roughly two-thirds 800 barrels of this volume was recovered.134

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines2.000.008.006.000.1L1.0 10.0 100.0 Spill Size Barrels1,000.0 10,000.0 100,000.0- California Regulated Hazardous Liquid Pipelines- California Crude Oil Pipelines Under Study Table 6-IA Spill Size Distribution Regulated California Haz. Liquid versus Crude Oil Pipelines Under Study Spill Size versus Incident Rate - Logarithmic Scale ~ 8.00 Co a, >~ a, 6.00 0 0 0 : 4.00 a, Q. U, a, C C a, CO C C C U, I- Co a, >. a, E 0 0 0 ` 4.00 a C. U' C a, 2.00 C C ~0.0o CO 0 C a, V C C- Spill Size DistributionSpill Size versus Cumulative Percentage of Incidents 0 to 100 Barrels Only20 40 Spill Size Barrels- California Regulated Hazardous Liquid Pipelines- California Crude Oil Pipelines Under Study60 80 100135

STAThRE MAZSHAL California State Fire Marshal JuLy 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gaxhermg Lines Table 6-lB Property Damage Distribution Regulated California Hazardous Liquid Pipelines versus Crude Oil Pipelines Under Study Incident Rate - Logarithmic Scale - California Regulated Hazardous Liquid Pipelines - California Crude Oil Pipelines Under Study Damage Distribution Logarithmic Scale - California Regulated Hazardous Liquid Pipelines - California Crude Oil Pipelines Under Study1,000 10.000 100,000 1,000,000 Damage $US 1994 8.00a,>.U, 6.00000: 4.00U,0~ 2.000C.~ 0.00a,~ 100CU,-o0Ci~ 8.00I.-a,>~a 6.00000: 4.00aICU 2.00-CCC.~ 0.00Co~ 100Ca,*0C,C10,000,0001,000 10,000Damage $US 1994100,000136

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~r~MARS~LAL Injuries and Fatalities No injuries or fatalities occurred on the California crude oil pipelines under study during the three year study period. However, the data sample was too small to be useful. For example, if one simply applied the fatality rate or 0.042 fatalities per 1,000 mile years, which resulted from the regulated California hazardous liquid pipelines, one would anticipate a fatality every 16 years for the 496 miles of crude oil pipelines under study. This recurrence interval is greater than the three year study period. As a result, one would not expect a fatality during this study. Further, as discussed above, the risk to human life from crude oil spills is likely less than for refined petroleum product pipelines, which would tend to increase the recurrence interval. 6.2 Incentive Option Investigation Findings After compiling all of the study information on incentive options and barriers to pipeline replacement and/or improvement, a number of conclusions or findings can be drawn. This subsection summarizes these fmdings and provides recommendations to improve pipeline safety public safety and environmental protection, maintain adequate regulatory control, and allow pipeline operators to make sound business/economic decisions. Most findings presented in this section were taken directly from responses to the questionnaire and from the case studies that were submitted. As noted above, the rate of response to the battery of questions on barriers was relatively low for the participating regulatory agencies. Therefore, it is important to remember that the findings and recommendations presented here do not necessarily reflect those that would have been obtained if a larger number of regulators had provided input. The major findings are summarized below: jurisdictional authority is not well defined, permitting requirements overlap, there is no lead agency for permitting, compliance requirements vary from agency to agency and from location to location, permitting process is often too slow, some permits require overly burdensome testing, some pipeline repair and replacement projects, including routine maintenance, are not being done, and 137

sTA~ra!!~ MA*SHAI. California State Fire Marshal July 1996 Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines incentives to repair, replace and improve pipelines do not exist or have proven ineffective.The following table lists some of the regulatory agencies involved in pipeline. StateJurisdiction LocalJurisdiction FederalJurisdictionCalifornia Coastal CommissionResource Management DepartmentMinerals Management ServiceState Lands CommissionPublic Works DepartmentDepartment of TransportationDepartment of Parks and RecreationFire DepartmentEnvironmental Protection AgencyWater Quality Control BoardEnvironmental Health ServicesUnited States Coast GuardState Fire MarshalBoard of Architectural ReviewNational Marine FisheriesDepartment of Fish and GameAir Pollution Control DistrictArmy Corps of EngineersCalifornia Air Resources BoardSystems Safety andReliability Review CommitteeFish and Wildlife ServiceZoning_DepartmentCalifornia Department of TransportationPlanning Department Bureau of LandManagement BLMBuilding Department138

California State Fire MarshalJuly 1996 ~Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA~.!!~MA~SHAL7.0 Recommendations 7.1 Database and Study Although the data set for the California crude oil pipelines under study was relatively small, it was sufficient to determine an overall leak incident rate. This incident rate was essentially the same as the incident rate for regulated California hazardous liquid pipelines. Although the overall leak incident rates for these groups of pipelines were similar, the likelihood of large spills, and spills resulting in large values of damage, were much lower for the crude oil pipelines under study. And finally, although the data was limited, there was no evidence to suggest that crude oil spills pose a significant risk to human life. As a result, we recommend that the California Government Code be modified as follows: Develop a set of criteria which can be used to identify pipelines which would likely impact unusually sensitive areas in the event of a leak. These criteria might include: likelihood of a spill from a given pipeline to reach a stream or waterway; etc. Distribute this criteria to the owners of the pipelines identified in this study. The operators would then identify those pipelines which would likely impact unusually sensitive areas in the event of a leak. Include the pipelines identified which would likely impact unusually sensitive areas into the existing regulations regarding intrastate hazardous liquid pipelines. Modify the law to require leak and pipeline inventory reporting for these lines, using the forms developed for this study. This will enable the CSFM to keep the database current. In addition to these statutory recommendations, we recommend the following actions: Continue to invite the operators of these pipelines to attend the regular Safety Seminars and other training programs provided by the CSFM. These programs would also be useful to other local and state public agencies. The database and mapping effort conducted as part of this study should be expanded to include California's intrastate and interstate pipelines. Funding should be appropriated to support this effort. An abbreviated report, covering the items included in Section 4.0 of this study should be prepared every 5 to 10 years. The goals of this study should be to identify incident rate trends, review current regulatory effectiveness, and recommend change. 139

Pr j1~ California State Fire Marshal July1996STATEFIREMA~S11AL Risk Assessment of Califoi-nia Low Pressure Crude Oil and Crude Oil Gathering Lines The permitting process for pipeline replacement or upgrade projects should be streamlined to the greatest extend possible. See also Section 7.2 of this report. Operators should be encouraged to replace older pipelines, when appropriate, to ensure pipeline safety. 7.2 Barriers and Incentive Options The State of California has clearly made a number of strides toward clarifying its jurisdictional authority over oil and gas transportation facilities - most notably in Section 51015.05 of the California Government Code. This 1994 legislation, by defining operative terms such as "production tanks and facilities" and "transportation facilities," resolved confusion and clearly distinguished between the jurisdictional authority of the CSFM and that of the DOGGR. In addition, this law is the driving force behind this study of incentive options and barriers to pipeline replacement and/or improvement in California. As possible evidence of the success of this statute, there was no indication by participants in this study that there is any lingering conflict between the jurisdictional responsibilities of DOGGR and the CSFM. Nevertheless, this study identified a number of levels of jurisdictional conflict and confusion. Although there was no evidence of perceived conflict among State-level agencies, it is clear that operators in particular perceive a tremendous amount of conflict between State-level agencies, on the one hand, and federal, county, and city agencies on the other. One of the most striking conclusions therefore, is that the perception of problems appears to be a serious problem for the State of California. Although the scope of this study particularly the questionnaire did not provide for independent verification or critical analysis of the information provided by the respondents, it is clear that there are any number of perceived barriers to pipeline replacement and improvements - these perceived barriers are particularly acute at the local level. Although detailed recommendations and specific implementation plans would be premature at this time, a number of general suggestions can be made. These suggestions should provide a useful backdrop and help guide the State of California as it further investigates its permitting process. The State should appoint a single lead agency with jurisdiction over every aspect of the permitting process in California. This lead agency should work in a partnering relationship between State and local agencies, with consideration for local land use and other issues. One of the agency's objectives should be to integrate federal, state and local policies for crude oil production and transportation. All permitting requirements should be standardized and redundancies and conflicts should be eliminated. A rigorous evaluation of the permitting process should be undertaken by the140

California State Fire Marshal ~July 1996 I~JRisk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines STA ~1REM4~SKAL new lead agency. Each requirement should be justified using sound scientific or other compelling reasoning. The lead agency should develop and implement a time line for permit application and approval. This time line should include "consequences" for the agency or operator for not meeting scheduled milestones. The lead agency should consider the following incentives to repair, replace, or improve pipelines: exclusion from CEQA requirements; reduction in the frequency of inspections for new pipelines; reduction of hydrostatic test frequency; etc. The most obvious incentive for operators to improve, repair or replace pipelines will be the comprehensive streamlining of state and local regulations. 141

cf'~g. California State Fire Marshal July 1996srF~tkgsH*l. Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines8.0 Bibliography CONCAWE Oil Pipelines Management Group's Special Task Force on Pipeline Spillages OP/STF-1. Performance of Oil Industry Cross Country Pipelines in Western Europe. Statistical Summary of Reported Spillages. 1981 to 1994 annual reports. Jones, Dana J., et. a!. An Analysis of Reportable Incidents for Natural Gas Transmission and Gathering Lines 1970 through June 1984. American Gas Association. March 1986. Jones, Dana J. and R. J. Eiber. An Analysis of Reportable Incidents for Natural Gas Transmission and Gathering Lines June 1984 through 1989. American Gas Association, October 1990. Line Pipe Research Supervisory Committee of the Pipeline Research Committee of the American Gas Association. An Analysis of Reportable Incidents for Natural Gas Transmission and Gathering Lines 1970 Through June 1984, NG- 18 Report Number 158. 1989. Line Pipe Research Supervisory Committee of the Pipeline Research Committee of the American Gas Association. An Analysis of DOT Reportable Incidents for Gas Transmission and Gathering Pipelines for June 1984 Through 1992, NG- 18 Report Number 213. 1995. Lyons, D., et. al. Performance of Oil Industry Cross Country Pipelines in Western Europe. Statistical Summary of Reported Spillages - 1981 through 1989. Brussels: CONCAWE. Martinsen, W. E. and J. B. Cornwell. "Use and Misuse of Historical Pipeline Failure Rate Data." Proceedings of the 1991 Pipeline Risk Assessment. Rehabilitation and Repair Conference, Gulf Publishing, 1991. National Transportation Safety Board. Pipeline Accident Report. Four Corners Pipe Line Company Rupture and Fire. Long Beach. California. December 1. 1980. August 1981. National Transportation Safety Board. Railroad Accident Report. Derailment of Southern Pacific Transportation Company Freight Train on May 12. 1989 and Subsequent Rupture of Calnev Petroleum Pipeline on May 25. 1989- San Bernardino. California. June 1990. Payne, Brian L., et. al. Hazardous Liquid Pipeline Risk Assessment, California State Fire Marshal, March 1993. Payne, Brian L., "California Hazardous Liquid Pipeline Risk Assessment." Proceedings of the 1993 International Pipeline Risk Assessment. Rehabilitation and Repair Conference, Gulf Publishing, September 1993. Payne, Brian L., "Block Valve Effectiveness and Seismic Activity - California Hazardous Liquid Pipelines." Proceedings of the 1994 International Pipeline Risk Assessment. Rehabilitation and Repair Conference, Gulf Publishing, September 1994. 142

California State Fire MarshalJuly 1996Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Sm FIRE MI~SHAL Payne, Brian L., "California Hazardous Liquid Pipeline Risk Assessment." Proceedings of the 1994 API Pipeline Conference, American Petroleum Institute, April 1994. United States Department of Transportation, Research and Special Programs Administration, Office of Pipeline Safety. Annual Report on Pipeline Safety. 1986 through 1992 annual reports. United States Department of Transportation, Research and Special Programs Administration. 1995 National Transportation Statistics - Annual Report. 143

California State Fire Marshal July 1996STA~!!R~M.~SHAL Risk Assessment of California Low Pressure Crude Oil and Crude Oil Gathering Lines Exhibit 1 AcknowledgmentsWe'd like to thank the following individuals and their companies for participating in this endeavorand providing the data necessary to conduct this study. Michael R. Madden Michael Lewis All American Pipeline Santa Fe Energy Resources Robert Anderson L. W. Alexander Arco Pipe Line Company Shell Pipe Line Sam Garland Troy Valenzuela Astarta Oil Company Stocker Resources Charles Dobie Dirk Cavanaugh Crutcher-Tufts Corporation Texaco Trading & Transportation, Inc. Milo Soto Gary M. Green Chemoil Refining Corporation Unocal Corporation M. A. Dirks Barry D. Emeneger Chevron Pipeline Unocal Corporation Michael L. Redfem Rod Eson G.E.N.Y. Operations V9noco Lynn Hall Joey A. Barulich H. L. Hall and Sons Vintage Petroleum, Inc. Ron Michaels Charles Lankford Koch Oil Witco Paul Strickland William B. Morrow Medallion California Wolf Petroleum, Inc. Mick Grocott Mobil Oil 144

Date post:	14-Jul-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times